For cafes, restaurants, bakeries, and food trucks: turns supplier prices, yields, and recipes into exact cost per portion, food cost percent on the tax-free price, contribution margin, and a suggested price, then applies menu engineering (Star, Plowhorse, Puzzle, Dog) with a tested Python calculator.
---
name: menu-food-cost-calculator
description: Costs recipes and menu items for cafes, restaurants, bakeries, food trucks, and caterers - converts purchase prices and yields into an exact cost per portion, food cost percentage on the tax-free price, contribution margin, and a suggested price at a target food cost, then classifies items with menu engineering (Star, Plowhorse, Puzzle, Dog) and recommends price, portion, and menu changes. Use when a user asks "what does this dish cost me?", "how should I price my menu?", "why is my food cost so high?", or shares recipes with supplier prices.
---
# Menu Food Cost and Pricing Calculator
You help small food businesses know what every plate really costs and price it with confidence. You work from real purchase prices and recipes, you show the math, and you think about margin in money, not only in percentages.
## Files in this skill
- `scripts/cost_menu.py` - costs every recipe from a JSON costing sheet, suggests prices, and runs menu engineering (Python 3 standard library only)
- `references/food-cost-basics.md` - yield, as-purchased versus edible cost, food cost percent, taxes, and common costing mistakes
- `references/pricing-strategies.md` - target-percent pricing, margin-based pricing, rounding, and menu engineering actions
- `templates/recipe-costing-sheet.md` - the JSON costing sheet the script reads, plus the report layout
- `examples/example-cafe-menu.md` - a worked review of a five-item cafe menu
## Workflow
### 1. Collect the inputs
Ask for or confirm:
- Currency, and whether menu prices include VAT or sales tax (and the rate).
- Target food cost percent (typical ranges are in `references/food-cost-basics.md`; default 30).
- For each ingredient: purchase price, pack size and unit, and yield (usable share after trimming, peeling, cooking loss, or spoilage).
- For each item: recipe quantities as prepared amounts, number of portions per batch, current menu price, packaging or garnish per portion, and weekly sales if known.
If something is missing, use a clearly labeled assumption (for example "yield 90 percent assumed for avocados") and list it in the report.
### 2. Build the costing sheet
Fill in `templates/recipe-costing-sheet.md` as JSON. Use units the script knows (g, kg, ml, l, oz, lb, each). If an ingredient is bought by the piece but used by weight, weigh one piece and convert; never mix dimensions.
### 3. Run the calculator
```bash
python3 scripts/cost_menu.py menu.json
python3 scripts/cost_menu.py menu.json --target 28
python3 scripts/cost_menu.py menu.json --json
```
The table shows cost per portion, menu price, net price without tax, food cost percent, contribution margin (net price minus cost), the price at the target food cost (rounded up), and the menu engineering class when weekly sales are given for every item. Errors (unknown ingredients, unit mismatches) and HIGH findings make the exit code 1.
If you cannot run the script, do the same calculation by hand, line by line, and say so.
### 4. Recommend
Use `references/pricing-strategies.md`:
1. Fix data errors first and rerun.
2. For HIGH and WARN items choose between raising the price, trimming the portion, changing an expensive ingredient, or accepting a higher percent because the money margin is strong. Name the trade-off.
3. Use the menu engineering class to decide where an item belongs on the menu and whether to promote, reprice, rework, or remove it.
4. Rerun with the proposed changes to show the before and after.
### 5. Report
Use the report layout in `templates/recipe-costing-sheet.md`, as in `examples/example-cafe-menu.md`.
## Rules
- Show the formula for at least one item so the owner can check it: cost per portion / net price x 100.
- Never treat the price at target as an instruction to lower an existing price; it is a benchmark.
- Do not give tax or legal advice; only apply the tax rate the user provides.
- Respect allergens and dietary claims when suggesting substitutions, and never suggest lowering food safety or quality standards.
- Recheck costs when supplier prices change by more than about 5 percent.
FILE:references/food-cost-basics.md
# Food cost basics
## Key terms
- **As-purchased (AP) cost**: what you pay for the pack, case, or piece.
- **Yield percent**: the usable share after trimming, peeling, deboning, cooking loss, or spoilage. Salmon fillet trimmed of skin and pin bones might yield 85 percent; whole avocados where 1 in 10 is unusable yield 90 percent when counted by the piece.
- **Edible portion (EP) cost** = AP cost per unit / (yield percent / 100). Recipes list prepared, usable quantities, so they are costed at EP cost.
- **Plate cost (cost per portion)** = sum of ingredient EP costs for the batch / portions + extras per portion (packaging, napkin, garnish, sauce cup).
- **Net price** = menu price / (1 + tax rate), when menu prices include VAT or sales tax. Food cost must be measured against the money you keep, not the tax you collect.
- **Food cost percent** = plate cost / net price x 100.
- **Contribution margin** = net price - plate cost. This is the money each sale leaves to pay labor, rent, and profit.
## Worked formula
Salmon fillet bought at 32.00 per kg with 85 percent yield:
- AP cost per g = 32.00 / 1000 = 0.032
- EP cost per g = 0.032 / 0.85 = 0.0376
- 160 g portion = 160 x 0.0376 = 6.02
## Typical food cost targets (rough guide)
| Concept | Typical food cost percent |
| --- | --- |
| Coffee and espresso drinks | 15 to 25 |
| Bakery items | 20 to 30 |
| Cafe brunch dishes | 28 to 35 |
| Casual restaurant mains | 28 to 35 |
| Steak and seafood mains | 35 to 45 |
| Pizza | 20 to 28 |
| Catering trays | 25 to 35 |
Your right target depends on labor, rent, and volume. A low-labor item can run a higher food cost percent and still be very profitable.
## Common costing mistakes
1. Forgetting small items: oil, butter for the pan, salt, garnish, sauces, and takeaway packaging. Add them or use extras_per_portion.
2. Using AP cost without yield for proteins and produce.
3. Measuring food cost against prices that include tax.
4. Costing the recipe card instead of what the kitchen actually plates (portion creep). Weigh five real portions.
5. Old supplier prices. Update the sheet when a price moves by about 5 percent or more.
6. Mixing units: an ingredient bought by the piece but used by weight needs one piece weighed.
7. Ignoring waste and staff meals; track them separately and compare actual food cost (from inventory) with this theoretical cost.
## Theoretical versus actual food cost
This skill calculates theoretical cost: what food should cost if recipes are followed. Actual food cost = (opening inventory + purchases - closing inventory) / net food sales. A gap of more than about 2 to 3 points usually means waste, portion creep, theft, or wrong prices on the sheet.
FILE:references/pricing-strategies.md
# Pricing strategies and menu engineering
## Ways to set a price
1. **Target food cost percent**: price = plate cost / target x (1 + tax rate), rounded up. Simple, and the script's "AT TARGET" column. Weak spot: cheap items end up underpriced and expensive proteins overpriced.
2. **Contribution margin**: decide the money each item must earn (for example at least 5.00 for a main), then price = (plate cost + margin) x (1 + tax rate). Better for high-cost proteins.
3. **Market check**: compare with three to five similar places nearby. Price perception matters as much as cost.
4. **Blend**: start from the target price, check the margin in money, then sanity-check against the market.
## Rounding and presentation
- Round up to the step your menu uses (0.10, 0.50, or whole numbers). Upscale menus often use whole numbers without currency signs; casual menus often end in .50 or .90.
- Avoid many small increases across the whole menu at once; raise the items with the weakest margin first.
- Keep price gaps logical: an oat milk upgrade should cover its extra cost (oat drink often costs about twice as much as dairy milk per liter).
## Menu engineering
Needs weekly sales for every item. The script uses:
- **Popularity line**: an item is popular if it sells at least 70 percent of an equal share (with 5 items, 0.7 x 20 percent = 14 percent of units sold).
- **Margin line**: the sales-weighted average contribution margin.
| Class | Popularity | Margin | What to do |
| --- | --- | --- | --- |
| Star | high | high | Keep quality and portion consistent, place it in the best menu spot, small price increases are usually safe. |
| Plowhorse | high | low | Raise price a little, trim cost (portion, garnish, supplier), or pair it with a high-margin add-on. Do not remove it. |
| Puzzle | low | high | Promote it: better menu placement, a description, staff recommendation, a photo. Check the price is not scaring guests. |
| Dog | low | low | Rework the recipe or price, or remove it, unless it serves a purpose (kids menu, dietary option, signature item). |
## Choosing a fix for a high food cost item
| Option | Good when | Risk |
| --- | --- | --- |
| Raise the price | the item is popular and the market allows it | fewer sales if the jump is large |
| Trim the portion | portions are larger than guests expect | guests notice; keep value perception |
| Swap an ingredient | a cheaper equal-quality option exists | allergen and taste changes; update the menu text |
| Accept a higher percent | the money margin is the highest on the menu | needs volume to pay off |
| Remove the item | it is a Dog with no strategic role | regulars may miss it |
Always rerun the calculator with the proposed change and show before and after.
FILE:templates/recipe-costing-sheet.md
# Recipe costing sheet (input for scripts/cost_menu.py)
Save as `menu.json`. Quantities in recipes are prepared (usable) amounts.
```json
{
"currency": "EUR",
"target_food_cost_pct": 30,
"menu_price_includes_tax_pct": 10,
"price_rounding": 0.10,
"ingredients": [
{"name": "flour", "price": 0.95, "per": "1 kg"},
{"name": "butter", "price": 9.80, "per": "1 kg"},
{"name": "eggs", "price": 3.60, "per": "12 each"},
{"name": "blueberries", "price": 16.00, "per": "1 kg", "yield_pct": 95}
],
"recipes": [
{
"name": "Blueberry Muffin",
"portions": 12,
"menu_price": 3.20,
"sold_per_week": 90,
"extras_per_portion": 0.06,
"items": [["flour", "500 g"], ["butter", "180 g"], ["eggs", "3 each"], ["blueberries", "300 g"]]
}
]
}
```
Field notes:
- `per`: the pack you buy, as "<amount> <unit>" (g, kg, mg, ml, cl, dl, l, oz, lb, each).
- `yield_pct`: 1 to 100, default 100.
- `menu_price_includes_tax_pct`: 0 if menu prices are shown without tax.
- `sold_per_week`: give it for every item (or none) to get menu engineering classes.
- `extras_per_portion`: packaging, napkins, garnish, sauce cups, in money.
Run: `python3 scripts/cost_menu.py menu.json [--target 30] [--json]`
---
# Menu costing report: <business> - <date>
**Target food cost:** <x>% **Prices include tax:** <rate or no> **Currency:** <code>
**Assumptions:** <yields, missing prices, portion weights>
## Results (before)
| Item | Cost/portion | Price | Net | Food % | Margin | At target | Class |
| --- | --- | --- | --- | --- | --- | --- | --- |
## Formula check
<one item worked out line by line>
## What needs attention
1. **<item>** - <finding>. Options: <price / portion / ingredient / accept>. Recommendation: <one>.
## Proposed changes and results (after)
<changes, then the new table or the changed rows>
## Menu engineering actions
- Stars: <items and action>
- Plowhorses: <items and action>
- Puzzles: <items and action>
- Dogs: <items and action>
## Next steps
- <weigh real portions, update supplier prices, track actual food cost monthly>
FILE:examples/example-cafe-menu.md
# Example: a five-item cafe menu
**User:** "We are a small brunch cafe. Prices include 10 percent VAT and I want about 30 percent food cost. Here are my supplier prices and recipes. Why is my margin so thin?"
The sheet has 16 ingredients and 5 items with weekly sales (avocado yield 90 percent because about 1 in 10 is unusable; salmon 85 percent after trimming).
**Command:**
```bash
python3 scripts/cost_menu.py cafe-menu.json
```
**Output (before):**
```
ITEM COST PRICE NET FOOD% MARGIN AT TARGET CLASS
Avocado Toast 3.13 9.50 8.64 36.2 5.51 11.50 Star
Salmon Spinach Bowl 8.02 14.50 13.18 60.8 5.16 29.50 Puzzle
Flat White 0.72 3.80 3.45 21.0 2.73 2.70 Plowhorse
Oat Flat White 0.91 4.20 3.82 23.9 2.91 3.40 Plowhorse
Blueberry Muffin 0.79 3.20 2.91 27.2 2.12 3.00 Dog
Findings (8):
[HIGH] Salmon Spinach Bowl: food cost 60.8% is far above the 30% target; price EUR 29.50 or cut cost 4.06 per portion
[WARN] Avocado Toast: food cost 36.2% is above the 30% target; price at target would be EUR 11.50
[INFO] Salmon Spinach Bowl: salmon fillet is 76% of the cost; its price or portion matters most
[INFO] menu engineering: weighted average margin EUR 3.22, popularity line 14.0% of items sold
[INFO] ingredient 'truffle oil' is not used in any recipe
```
---
# Menu costing report: brunch cafe - October
**Target food cost:** 30% **Prices include tax:** 10% VAT **Currency:** EUR
**Assumptions:** avocado yield 90%, salmon 85%, spinach 90%; extras 0.10 per dish, 0.12 per coffee (cup and lid), 0.06 per muffin.
## Formula check (Salmon Spinach Bowl, before)
- Salmon 160 g x (32.00 / 1000 / 0.85) = 6.02
- Spinach 70 g x (14.00 / 1000 / 0.90) = 1.09; egg 0.30; tomatoes 0.36; olive oil 0.15; extras 0.10
- Cost per portion = 8.02; net price = 14.50 / 1.10 = 13.18; food cost = 8.02 / 13.18 x 100 = 60.8%
## What needs attention
1. **Salmon Spinach Bowl (HIGH, Puzzle)** - 60.8% food cost; salmon is 76% of the cost. Pricing it at target (29.50) is unrealistic for a cafe. Recommendation: reduce salmon to 120 g (still a generous portion for a bowl) and raise the price to 17.50; accept about 40% food cost because the margin becomes the highest on the menu.
2. **Avocado Toast (WARN, Star)** - 36.2%. It is the best-selling dish, so a 1.00 increase to 10.50 is low risk.
3. **Flat White and Oat Flat White (Plowhorses)** - healthy percentages (21 to 24%) but small margins; do not discount. Keep the oat surcharge at 0.40: the oat drink costs 0.19 more per cup than milk.
4. **Blueberry Muffin (Dog)** - fine percentage, low margin and low sales. Try a bundle with coffee before removing it.
5. **Truffle oil** is on the sheet but in no recipe: remove it from orders or the sheet.
## Proposed changes and results (after)
Avocado Toast 10.50; Salmon Spinach Bowl 120 g salmon at 17.50. Rerun: `python3 scripts/cost_menu.py cafe-menu-revised.json`
```
ITEM COST PRICE NET FOOD% MARGIN AT TARGET CLASS
Avocado Toast 3.13 10.50 9.55 32.8 6.42 11.50 Star
Salmon Spinach Bowl 6.51 17.50 15.91 40.9 9.40 23.90 Puzzle
Flat White 0.72 3.80 3.45 21.0 2.73 2.70 Plowhorse
Oat Flat White 0.91 4.20 3.82 23.9 2.91 3.40 Plowhorse
Blueberry Muffin 0.79 3.20 2.91 27.2 2.12 3.00 Dog
Findings (6):
[WARN] Salmon Spinach Bowl: food cost 40.9% is above the 30% target; price at target would be EUR 23.90
[INFO] menu engineering: weighted average margin EUR 3.56, popularity line 14.0% of items sold
```
Exit code 0. The remaining WARN is accepted on purpose: 9.40 margin per bowl versus 5.16 before.
## Menu engineering actions
- Stars: Avocado Toast - keep the recipe consistent, top of the brunch section.
- Plowhorses: Flat White, Oat Flat White - no discounts; suggest a pastry with every coffee.
- Puzzles: Salmon Spinach Bowl - give it a short description and staff recommendation; check sales after 4 weeks at the new price.
- Dogs: Blueberry Muffin - test a coffee + muffin bundle for 4 weeks, then decide.
## Next steps
- Weigh five real salmon portions this week to confirm the 120 g spec is followed.
- Update supplier prices monthly and rerun the sheet.
- Compare with actual food cost from inventory at month end.
FILE:scripts/cost_menu.py
#!/usr/bin/env python3
"""Cost menu items from recipes and purchase prices, and suggest menu prices.
Usage:
python3 cost_menu.py menu.json [--target 30] [--json]
cat menu.json | python3 cost_menu.py -
Input JSON (see templates/recipe-costing-sheet.md):
{
"currency": "EUR",
"target_food_cost_pct": 30, # optional, default 30 (or --target)
"menu_price_includes_tax_pct": 10, # optional; VAT/sales tax included in menu prices
"price_rounding": 0.10, # optional; suggested prices round UP to this step
"ingredients": [
{"name": "butter", "price": 9.80, "per": "1 kg", "yield_pct": 100}
],
"recipes": [
{"name": "Croissant", "portions": 12, "menu_price": 3.20, "sold_per_week": 180,
"extras_per_portion": 0.05, # optional: packaging, napkin, garnish
"items": [["butter", "600 g"], ["flour", "1 kg"]]}
]
}
Units: g, kg, mg, ml, cl, dl, l, oz, lb, each (also pc, pcs, piece, unit, egg).
Recipe quantities are the prepared (usable) amounts. yield_pct is the usable
share of what you buy after trimming, peeling, cooking loss or spoilage.
Per recipe: cost per portion, food cost percent of the net (tax-free) menu
price, contribution margin, suggested price at the target, and the three
biggest cost drivers. With sold_per_week on every recipe, adds a menu
engineering class (Star, Plowhorse, Puzzle, Dog).
Exit code: 0 ok, 1 errors in the data or HIGH findings, 2 usage or input error.
Standard library only.
"""
import json
import math
import re
import sys
UNITS = { # unit -> (dimension, factor to base unit g / ml / each)
"mg": ("mass", 0.001), "g": ("mass", 1.0), "kg": ("mass", 1000.0),
"oz": ("mass", 28.3495), "lb": ("mass", 453.592),
"ml": ("volume", 1.0), "cl": ("volume", 10.0), "dl": ("volume", 100.0), "l": ("volume", 1000.0),
"each": ("count", 1.0), "pc": ("count", 1.0), "pcs": ("count", 1.0), "piece": ("count", 1.0),
"pieces": ("count", 1.0), "unit": ("count", 1.0), "units": ("count", 1.0), "egg": ("count", 1.0), "eggs": ("count", 1.0),
}
BASE = {"mass": "g", "volume": "ml", "count": "each"}
def usage(msg):
print(f"error: {msg}\n", file=sys.stderr)
print(__doc__.strip().split("\n\n")[1], file=sys.stderr)
sys.exit(2)
def parse_qty(text):
"""'600 g' -> (600.0, 'mass', 600.0 in base units)."""
m = re.fullmatch(r"\s*(\d+(?:[.,]\d+)?)\s*([a-zA-Z]+)?\s*", str(text))
if not m:
raise ValueError(f"cannot read quantity {text!r}")
qty = float(m.group(1).replace(",", "."))
unit = (m.group(2) or "each").lower()
if unit not in UNITS:
raise ValueError(f"unknown unit {unit!r} in {text!r}")
dim, factor = UNITS[unit]
return qty, dim, qty * factor
def round_up(value, step):
return math.ceil(round(value / step, 6)) * step
def analyze(data, target_override=None):
errors, findings = [], []
cur = data.get("currency", "")
target = float(target_override or data.get("target_food_cost_pct", 30))
tax = float(data.get("menu_price_includes_tax_pct", 0))
step = float(data.get("price_rounding", 0.10))
ingredients = {}
for ing in data.get("ingredients", []):
name = str(ing.get("name", "")).strip().lower()
try:
_, dim, base_qty = parse_qty(ing["per"])
price = float(ing["price"])
except (KeyError, ValueError, TypeError) as e:
errors.append(f"ingredient {name or '?'}: {e}")
continue
y = float(ing.get("yield_pct", 100))
if not 0 < y <= 100:
errors.append(f"ingredient {name}: yield_pct must be between 1 and 100")
continue
ingredients[name] = {"dim": dim, "cost_per_base": price / base_qty / (y / 100), "yield": y, "used": False}
results = []
for rec in data.get("recipes", []):
rname = rec.get("name", "?")
portions = float(rec.get("portions", 1) or 1)
lines, bad = [], False
for item in rec.get("items", []):
iname, qtext = str(item[0]).strip().lower(), item[1]
ing = ingredients.get(iname)
if not ing:
errors.append(f"{rname}: unknown ingredient {iname!r} (add it to ingredients)")
bad = True
continue
ing["used"] = True
try:
_, dim, base_qty = parse_qty(qtext)
except ValueError as e:
errors.append(f"{rname}: {e}")
bad = True
continue
if dim != ing["dim"]:
errors.append(f"{rname}: {iname} is bought by {BASE[ing['dim']]} but used by {BASE[dim]} ({qtext}); "
f"convert it (for example weigh one piece)")
bad = True
continue
lines.append((iname, base_qty * ing["cost_per_base"]))
if bad:
continue
batch = sum(c for _, c in lines)
extras = float(rec.get("extras_per_portion", 0))
cost = batch / portions + extras
price = float(rec.get("menu_price", 0))
net = price / (1 + tax / 100) if price else 0.0
pct = 100 * cost / net if net else None
suggested = round_up(cost / (target / 100) * (1 + tax / 100), step)
drivers = sorted(lines, key=lambda x: -x[1])[:3]
r = {"name": rname, "portions": portions, "cost_per_portion": round(cost, 3), "menu_price": price,
"net_price": round(net, 2), "food_cost_pct": None if pct is None else round(pct, 1),
"contribution_margin": round(net - cost, 2) if net else None,
"suggested_price_at_target": round(suggested, 2), "sold_per_week": rec.get("sold_per_week"),
"drivers": [{"ingredient": n, "share_pct": round(100 * c / batch, 1) if batch else 0} for n, c in drivers]}
results.append(r)
if pct is None:
findings.append(("WARN", rname, f"no menu_price; suggested {cur} {suggested:.2f} at {target:g}% food cost"))
elif pct > target + 15:
findings.append(("HIGH", rname, f"food cost {pct:.1f}% is far above the {target:g}% target; "
f"price {cur} {suggested:.2f} or cut cost {cost - net * target / 100:.2f} per portion"))
elif pct > target + 5:
findings.append(("WARN", rname, f"food cost {pct:.1f}% is above the {target:g}% target; "
f"price at target would be {cur} {suggested:.2f}"))
elif pct < target - 15:
findings.append(("INFO", rname, f"food cost only {pct:.1f}%; check the recipe lists every ingredient and portion size"))
if drivers and batch and drivers[0][1] / batch > 0.5:
findings.append(("INFO", rname, f"{drivers[0][0]} is {100 * drivers[0][1] / batch:.0f}% of the cost; "
f"its price or portion matters most"))
sold = [r for r in results if isinstance(r["sold_per_week"], (int, float)) and r["contribution_margin"] is not None]
if sold and len(sold) == len(results) and len(sold) >= 3:
total = sum(r["sold_per_week"] for r in sold)
pop_line = 0.7 / len(sold)
avg_cm = sum(r["contribution_margin"] * r["sold_per_week"] for r in sold) / total if total else 0
for r in sold:
high_pop = total and r["sold_per_week"] / total >= pop_line
high_cm = r["contribution_margin"] >= avg_cm
r["menu_class"] = {(True, True): "Star", (True, False): "Plowhorse",
(False, True): "Puzzle", (False, False): "Dog"}[(bool(high_pop), high_cm)]
findings.append(("INFO", None, f"menu engineering: weighted average margin {cur} {avg_cm:.2f}, "
f"popularity line {100 * pop_line:.1f}% of items sold"))
for name, ing in ingredients.items():
if not ing["used"]:
findings.append(("INFO", None, f"ingredient {name!r} is not used in any recipe"))
return {"currency": cur, "target_pct": target, "tax_pct": tax, "recipes": results,
"errors": errors, "findings": [{"severity": s, "recipe": n, "message": m} for s, n, m in findings]}
def main(argv):
target, as_json, paths = None, False, []
it = iter(argv)
for a in it:
if a == "--target":
try:
target = float(next(it, ""))
except ValueError:
usage("--target needs a number, for example 30")
if not 5 <= target <= 80:
usage("--target should be a food cost percent between 5 and 80")
elif a == "--json":
as_json = True
elif a.startswith("--"):
usage(f"unknown option {a}")
else:
paths.append(a)
if len(paths) != 1:
usage("give exactly one menu JSON file, or - for stdin")
try:
raw = sys.stdin.read() if paths[0] == "-" else open(paths[0], encoding="utf-8").read()
data = json.loads(raw)
except (OSError, ValueError) as e:
print(f"error: cannot read menu JSON: {e}", file=sys.stderr)
return 2
if not data.get("recipes"):
print("error: no recipes in the input", file=sys.stderr)
return 2
rep = analyze(data, target)
if as_json:
print(json.dumps(rep, indent=2))
else:
cur = rep["currency"]
print(f"Target food cost {rep['target_pct']:g}% | prices include {rep['tax_pct']:g}% tax | currency {cur}\n")
print(f"{'ITEM':<22} {'COST':>7} {'PRICE':>7} {'NET':>7} {'FOOD%':>6} {'MARGIN':>7} {'AT TARGET':>9} CLASS")
for r in rep["recipes"]:
pct = "-" if r["food_cost_pct"] is None else f"{r['food_cost_pct']:.1f}"
cm = "-" if r["contribution_margin"] is None else f"{r['contribution_margin']:.2f}"
print(f"{r['name'][:22]:<22} {r['cost_per_portion']:>7.2f} {r['menu_price']:>7.2f} {r['net_price']:>7.2f} "
f"{pct:>6} {cm:>7} {r['suggested_price_at_target']:>9.2f} {r.get('menu_class', '-')}")
print("\nTop cost drivers:")
for r in rep["recipes"]:
print(f" {r['name']}: " + ", ".join(f"{d['ingredient']} {d['share_pct']:g}%" for d in r["drivers"]))
if rep["errors"]:
print(f"\nErrors ({len(rep['errors'])}):")
for e in rep["errors"]:
print(f" [ERROR] {e}")
print(f"\nFindings ({len(rep['findings'])}):")
order = {"HIGH": 0, "WARN": 1, "INFO": 2}
for f in sorted(rep["findings"], key=lambda f: order[f["severity"]]):
print(f" [{f['severity']}] {f['recipe'] + ': ' if f['recipe'] else ''}{f['message']}")
bad = rep["errors"] or any(f["severity"] == "HIGH" for f in rep["findings"])
return 1 if bad else 0
if __name__ == "__main__":
sys.exit(main(sys.argv[1:]))Audits the licenses of your project's dependencies for how you actually ship (SaaS, distributed app or library, or internal tool): normalizes license names to SPDX, resolves OR and AND expressions, flags missing, AGPL, GPL, source-available, and incompatible licenses, and writes a ranked fix plan and notices list. Includes a tested stdlib Python auditor.
---
name: dependency-license-audit
description: Audits the licenses of your project's dependencies for how you actually ship (SaaS, distributed app or library, or internal tool): normalizes license names to SPDX, resolves OR and AND expressions, flags missing, AGPL, GPL, source-available, and incompatible licenses, and writes a ranked fix plan and notices list. Includes a tested stdlib Python auditor.
---
# Dependency License Audit
You help engineering teams find license problems in their dependencies before a release, an acquisition review, or a customer security questionnaire finds them. You turn a raw dependency list into a short, ranked list of things to fix, and you explain each one in plain language. You are a careful engineer, not a lawyer: you flag risk and say when a human with legal training must decide.
## Files in this skill
- `scripts/audit_licenses.py` - normalizes and classifies every dependency license and rates it for the chosen distribution model (Python 3 standard library only)
- `references/license-categories.md` - what permissive, weak copyleft, strong copyleft, network copyleft, and source-available mean in practice, with common SPDX ids
- `references/remediation-playbook.md` - how to fix each kind of finding, from replacing a package to adding notices
- `templates/license-audit-report.md` - the report layout
- `examples/example-saas-audit.md` - a worked audit of a Node.js SaaS backend
## Workflow
### 1. Understand how the project is shipped
Ask, or infer and state as an assumption:
- Project license: proprietary (closed source) or an open source license (SPDX id).
- Distribution: `saas` (users only reach it over a network), `distributed` (mobile or desktop apps, on-prem installs, devices, SDKs, published libraries, container images given to customers), or `internal` (only employees use it).
- Any company policy: always-allowed and always-blocked licenses.
The same dependency can be fine for one model and a blocker for another, so never skip this step.
### 2. Get the dependency inventory
Prefer a tool's output over a hand-made list, and include transitive dependencies:
- Python: `pip-licenses --format=json --with-system` (or `pip-licenses --format=csv`).
- Node.js: `npx license-checker --json --production` (run again without `--production` for dev tools).
- Anything else: a CSV with `name,version,license,scope` exported from an SBOM or the package manager.
Mark build, test, and lint tools as `scope=dev`; they are not shipped, so their findings drop to INFO.
### 3. Run the audit
```bash
python3 scripts/audit_licenses.py deps.csv --project-license proprietary --distribution saas
python3 scripts/audit_licenses.py licenses.json --distribution distributed --deny AGPL-3.0-only
python3 scripts/audit_licenses.py deps.csv --json
```
The script prints severity counts, a table of every package with its normalized license and category, findings with reasons, version-to-version license changes, and the licenses that need text in your third-party notices. Exit code 1 means at least one HIGH finding.
If you cannot run the script, classify each license by hand with `references/license-categories.md` and say so in the report.
### 4. Verify before you conclude
- For every HIGH or WARN finding, check the package's actual LICENSE file in its repository or package archive; metadata is often wrong or outdated.
- For "missing" or "SEE LICENSE IN" entries, open the referenced file and classify it by its text.
- For dual-licensed packages (OR), record which option you choose.
- For packages whose license changed between versions, confirm which version is actually installed (lock file).
### 5. Recommend and report
Use `references/remediation-playbook.md` to give one concrete action per finding: replace, isolate, pin an older permissive version, get a commercial license, comply (publish source or add notices), or accept with a recorded reason. Fill in `templates/license-audit-report.md`, as in `examples/example-saas-audit.md`.
## Rules
- Never say a license "is fine" or "is illegal" as a legal conclusion. Say what the license usually requires and when to ask legal counsel.
- Do not guess a license from a package name or popularity. Unknown stays unknown until someone reads the license text.
- Rate risk for the user's distribution model, and note what would change if the model changes (for example SaaS today, on-prem next year).
- Keep the report short: blockers first, then obligations (notices, source offers), then the clean list as a count.
- Treat the audit as a snapshot; recommend running it in CI on every dependency change.
FILE:references/license-categories.md
# License categories in practice
Severity in the script depends on the category and on how the project is shipped. This page explains the categories in plain language. It is an engineering summary, not legal advice.
## Public domain and near public domain
SPDX: `0BSD`, `Unlicense`, `CC0-1.0`, `MIT-0`, `WTFPL`
- No meaningful obligations. Keeping a notice is still good practice.
## Permissive
SPDX: `MIT`, `ISC`, `BSD-2-Clause`, `BSD-3-Clause`, `Apache-2.0`, `Zlib`, `BSL-1.0` (Boost), `PSF-2.0`, `Python-2.0`, `CC-BY-4.0` (data such as caniuse-lite)
- Use, modify, and ship in closed source products.
- Keep the copyright notice and license text, usually in a third-party notices file or an "About / Licenses" screen.
- Apache-2.0 also requires keeping any NOTICE file, stating significant changes to its files, and includes a patent grant that ends if you sue over patents in that code.
- BSD-3-Clause forbids using the authors' names to promote your product.
## Weak copyleft (file or library level)
SPDX: `LGPL-2.1-*`, `LGPL-3.0-*`, `MPL-2.0`, `EPL-1.0`, `EPL-2.0`, `CDDL-1.0`, `CDDL-1.1`, GPL with a linking exception (for example `GPL-2.0-only WITH Classpath-exception-2.0`)
- You may combine them with closed source code, but changes to the library's own files must be shared under the same license when you distribute.
- LGPL: users must be able to replace the library with their own version. Dynamic linking (shared library, separate jar, npm package) normally makes this easy; static linking or bundling into one binary needs care.
- MPL-2.0: copyleft applies per file. Keep MPL files separate and publish changes to those files.
## Strong copyleft
SPDX: `GPL-2.0-only`, `GPL-2.0-or-later`, `GPL-3.0-only`, `GPL-3.0-or-later`
- If you distribute software that includes or links GPL code, the combined work must be offered under the GPL with its source code.
- Running GPL software on your own servers and letting users reach it over a network is not distribution, so pure SaaS use is usually allowed. Shipping an app, an on-prem installer, a device firmware, or a container image to customers is distribution.
- Version matters: `GPL-2.0-only` code cannot be combined with GPL-3.0 or Apache-2.0 code in one work. `-or-later` gives more room.
## Network copyleft
SPDX: `AGPL-3.0-only`, `AGPL-3.0-or-later`, `SSPL-1.0` (not OSI approved)
- Like the GPL, plus: if users interact with a modified version over a network, you must offer them its source code.
- For a closed source SaaS product this is the most common blocker. Even unmodified use triggers internal reviews at most companies.
- SSPL goes further: offering the software as a service can require releasing the source of your whole service stack.
## Source-available and non-commercial
SPDX or names: `BUSL-1.1` (Business Source License), `Elastic-2.0`, Commons Clause add-ons, `CC-BY-NC-*`, `PolyForm-Noncommercial-1.0.0`
- Source is visible, but use is restricted: often no competing hosted service, no production use above a limit, or no commercial use at all, sometimes until a change date.
- Always needs a review of the exact terms. Many of these switch to an open license after a few years, so an older or newer version may differ.
## Unknown
- Empty license fields, `UNKNOWN`, `NOASSERTION`, `UNLICENSED`, `SEE LICENSE IN <file>`, `LicenseRef-*`, or a name the script does not recognize.
- "No license" means all rights reserved: you have no permission to use it. Find the real terms or replace the package.
## Expressions
- `MIT OR Apache-2.0`: you may choose either; pick the one that suits you and record the choice.
- `MIT AND BSD-3-Clause`: you must follow both; the strictest one sets the category.
- `GPL-2.0-only WITH <exception>`: the exception narrows the copyleft; read it.
FILE:references/remediation-playbook.md
# Remediation playbook
Pick one action per finding and write it down with an owner. Prefer the cheapest action that removes the risk.
## Order of work
1. Unknown and missing licenses (you may have no right to use the code at all).
2. Network copyleft and source-available packages in shipped or hosted code.
3. Strong copyleft in distributed products, and version incompatibilities in open source projects.
4. Weak copyleft obligations (linking, publishing changes).
5. Notices for everything else.
## Actions
### Find the real license
- Open the package's repository and archive, read LICENSE, COPYING, or the header of the main source file.
- Check whether the metadata is just stale (a common case for old packages and forks).
- If there is truly no license, ask the maintainer to add one (link to choosealicense.com in the request) or replace the package.
### Replace the package
- Search for a maintained alternative with a permissive license and similar API.
- Estimate effort in hours and add it to the backlog with the release it blocks.
- For transitive dependencies, look for the direct dependency that pulls it in; a different version or a configuration flag often drops it.
### Pin a version under the old license
- When a project switched from a permissive license to source-available (or the reverse), the old versions keep their old license.
- Pinning is a short-term fix: no security updates. Schedule a proper replacement.
### Isolate it
- Run the copyleft tool as a separate process or service that talks to your code over a command line, HTTP, or a queue, instead of linking it into your code.
- Keep it unmodified if possible. Document the boundary.
- This is a common way to use GPL command line tools, but have legal counsel confirm it for AGPL and SSPL.
### Buy a commercial license
- Many dual-licensed projects (open source plus commercial) sell a license that removes copyleft obligations. Note the price and renewal terms.
### Comply
- Weak copyleft: keep the library as a separate, replaceable file or package; publish any changes to its own files; include its license.
- GPL in a distributed product: only if the whole product can be open source under compatible terms. This is a business decision, not an engineering one.
- AGPL in a hosted service: offer the corresponding source to users of the service, including your changes.
### Accept with a recorded reason
- Dev-only tools (linters, test runners, build tools) that never ship are usually fine. Record "dev only, not distributed".
- Internal-only use of GPL or AGPL tools is usually fine. Record who uses it and that it is not exposed to outside users.
## Third-party notices
- Generate a notices file from the inventory: package, version, license, copyright line, full license text (one copy per distinct license text).
- Include Apache NOTICE files as they are.
- Ship it inside apps ("Settings > About > Licenses"), installers, and container images; for SaaS, link it from the footer or docs.
## Keep it clean
- Run the audit in CI on every change to a lock file and fail the build on HIGH.
- Keep an allow list (for example MIT, BSD, ISC, Apache-2.0) and a deny list (for example AGPL-3.0, SSPL-1.0, BUSL-1.1) agreed with legal counsel, and pass them with `--allow` and `--deny`.
FILE:templates/license-audit-report.md
# Dependency License Audit: <project name>
Date: <YYYY-MM-DD>
Project license: <proprietary | SPDX id>
Distribution: <saas | distributed | internal> (<one line: how users get the software>)
Inventory source: <tool and command, lock file commit>
Packages audited: <n> (<runtime n>, <dev n>)
## Summary
- Blockers (HIGH): <n> - <one line naming them>
- Obligations (WARN): <n> - <one line>
- Notes (INFO): <n>
- Clean (OK): <n>
- Verdict: <ready to ship | ship after fixes | blocked> because <reason>.
## Blockers
| Package | Version | License (as found) | Category | Why it matters for us | Action | Owner | Due |
|---|---|---|---|---|---|---|---|
| | | | | | | | |
## Obligations
| Package | License | What we must do | Done? |
|---|---|---|---|
| | | | |
## Verified by hand
- <package>: metadata said <x>, LICENSE file says <y>.
## Decisions and assumptions
- Dual-licensed packages and the option chosen: <package: option>
- Accepted risks with reason: <package: reason, approved by>
- Assumptions: <for example "the admin dashboard is internal only">
## What changes if we change how we ship
- <for example: "If we ship an on-prem version, readline-sync (GPL-3.0) becomes a blocker.">
## Third-party notices
Licenses that need full text in the notices file: <list>
Notices file location: <path or URL>
## Next audit
Run in CI on every lock file change; full review before <next release or date>.
_This report is an engineering triage, not legal advice. Blockers must be confirmed by legal counsel before relying on any exception._
FILE:examples/example-saas-audit.md
# Example: license audit of a Node.js SaaS backend
## The request
"We're a closed source B2B SaaS. Enterprise customers keep sending security questionnaires asking about open source licenses, and next year we may sell an on-prem version. Here is our dependency export. Anything we should worry about?"
Inventory (`deps.csv`, exported from the lock file, 18 rows):
```
name,version,license,scope
express,4.19.2,MIT,runtime
lodash,4.17.21,MIT,runtime
pg,8.12.0,MIT,runtime
sharp,0.33.4,Apache-2.0,runtime
dompurify,3.1.6,(MPL-2.0 OR Apache-2.0),runtime
ghostscript-wrapper,2.1.0,AGPL-3.0-only,runtime
chart-kit-pro,1.4.0,SEE LICENSE IN LICENSE.md,runtime
readline-sync,1.4.10,GPL-3.0,runtime
mariadb-connector,3.3.1,LGPL-2.1-or-later,runtime
fast-csv-lite,0.9.0,,runtime
redis-om,0.4.3,Business Source License 1.1,runtime
jest,29.7.0,MIT,dev
eslint-plugin-gpl-thing,1.0.2,GPL-2.0-only,dev
argparse,2.0.1,Python-2.0,runtime
tslib,2.6.3,0BSD,runtime
lodash,3.10.1,MIT,runtime
search-client,7.10.2,Apache-2.0,runtime
search-client,8.1.0,Elastic License 2.0,runtime
```
## Script run
```
$ python3 scripts/audit_licenses.py deps.csv --project-license proprietary --distribution saas
Dependency license audit: 18 packages, project license proprietary, distribution saas
HIGH 5 WARN 1 INFO 2 OK 10
...
HIGH chart-kit-pro@1.4.0 SEE LICENSE IN LICENSE.md unknown runtime
HIGH fast-csv-lite@0.9.0 (missing) unknown runtime
HIGH redis-om@0.4.3 BUSL-1.1 source-available runtime
HIGH search-client@8.1.0 Elastic-2.0 source-available runtime
HIGH ghostscript-wrapper@2.1.0 AGPL-3.0-only network-copyleft runtime
WARN readline-sync@1.4.10 GPL-3.0-only strong-copyleft runtime
INFO eslint-plugin-gpl-thing@1.0.2 GPL-2.0-only strong-copyleft dev
INFO mariadb-connector@3.3.1 LGPL-2.1-or-later weak-copyleft runtime
...
Notes:
- search-client: different licenses across versions (Apache-2.0, Elastic-2.0); check which version you ship.
```
Exit code 1 (HIGH findings).
## The report (filled template, shortened)
# Dependency License Audit: acme-api
Project license: proprietary. Distribution: saas (customers use the web app and REST API; nothing is installed on their machines today).
Packages audited: 18 (16 runtime, 2 dev).
### Summary
- Blockers (HIGH): 5 - two packages with no usable license, two source-available packages, one AGPL package.
- Obligations (WARN): 1 - a GPL-3.0 package that is fine for SaaS but blocks the planned on-prem version.
- Verdict: ship after fixes. Nothing here is unusual, and each item has a cheap fix.
### Blockers
| Package | License (as found) | Why it matters for us | Action |
|---|---|---|---|
| chart-kit-pro 1.4.0 | SEE LICENSE IN LICENSE.md | The LICENSE.md is a commercial license that allows use only with a paid key. We have no key on record. | Confirm with the team that bought it, or replace with a permissive charting library. |
| fast-csv-lite 0.9.0 | (missing) | No license means no permission to use. | Ask the maintainer to add a license; meanwhile switch to a maintained CSV parser with MIT terms. |
| redis-om 0.4.3 | BUSL-1.1 | Business Source License limits production use until its change date; terms vary per project. | Read the exact "Additional Use Grant". If it forbids our use, replace or buy a license. |
| search-client 8.1.0 | Elastic-2.0 | Elastic License 2.0 forbids offering the software as a managed service. Version 7.10.2 in the same tree is Apache-2.0. | Check the lock file: which modules import 8.1.0? Pin both to the Apache-2.0 line or move to an Apache-licensed fork. |
| ghostscript-wrapper 2.1.0 | AGPL-3.0-only | We render customer PDFs with it inside the API; AGPL network use requires offering source to users. | Move PDF rendering to a separate, unmodified worker service and have legal counsel confirm, or buy the commercial license. |
### Obligations
| Package | License | What we must do |
|---|---|---|
| readline-sync 1.4.10 | GPL-3.0-only | Fine for SaaS. It becomes a blocker if the on-prem version ships with it. It is only used by an admin CLI script, so remove it from the server image. |
| mariadb-connector 3.3.1 | LGPL-2.1-or-later | Keep it as a normal npm dependency (replaceable); publish changes if we ever patch it. |
### Verified by hand
- dompurify: dual licensed `MPL-2.0 OR Apache-2.0`; we choose Apache-2.0.
- eslint-plugin-gpl-thing: dev only, never shipped. Accepted.
### What changes if we ship on-prem
- readline-sync (GPL-3.0) becomes HIGH, mariadb-connector (LGPL) needs the replaceability check, and Apache-2.0 packages need their NOTICE files in the installer. Rerun with `--distribution distributed` before that release.
### Third-party notices
Licenses that need full text: Apache-2.0, LGPL-2.1-or-later, MIT, Python-2.0. Link the notices page from the app footer.
_Engineering triage, not legal advice._
## Why this is a good answer
- It rates every package for this company's real distribution model, and shows what changes with the on-prem plan.
- It checks the facts the metadata cannot show (the commercial LICENSE.md, the duplicate search-client versions).
- Every blocker has one concrete, cheap action.
FILE:scripts/audit_licenses.py
#!/usr/bin/env python3
"""Audit third-party dependency licenses against how a project is shipped.
Usage:
python3 audit_licenses.py deps.csv --project-license proprietary --distribution saas
python3 audit_licenses.py licenses.json --distribution distributed [--json]
cat deps.csv | python3 audit_licenses.py - [--allow LIC,LIC] [--deny LIC,LIC]
Input (detected automatically):
CSV with a header: name,version,license[,scope] (scope: runtime | dev; default runtime)
JSON list: [{"name": "...", "version": "...", "license": "...", "scope": "dev"}]
pip-licenses --format=json: [{"Name": "...", "Version": "...", "License": "..."}]
license-checker --json (npm): {"pkg@1.2.3": {"licenses": "MIT"}, ...}
Options:
--project-license your project's license: proprietary (default) or an SPDX id such as MIT, Apache-2.0, GPL-3.0-only
--distribution saas (network service, default) | distributed (apps, binaries, devices, libraries) | internal
--allow / --deny comma-separated SPDX ids that your policy always allows or always blocks
--json machine-readable output
License names are normalized to SPDX ids, expressions with OR take the most
permissive option and AND takes the most restrictive one. Each package gets a
category and a severity (OK, INFO, WARN, HIGH) for the chosen distribution.
Dev-only packages are not shipped, so their findings drop to INFO.
Exit code: 0 no HIGH findings, 1 at least one HIGH, 2 usage or input error.
This is an engineering triage aid, not legal advice. Standard library only.
"""
import argparse
import csv
import io
import json
import re
import sys
ALIASES = { # lowercased common names -> SPDX id
"mit": "MIT", "mit license": "MIT", "the mit license": "MIT", "expat": "MIT",
"isc": "ISC", "isc license": "ISC", "isc license (iscl)": "ISC",
"apache 2.0": "Apache-2.0", "apache-2": "Apache-2.0", "apache 2": "Apache-2.0", "apache2": "Apache-2.0",
"apache license 2.0": "Apache-2.0", "apache license, version 2.0": "Apache-2.0",
"apache software license": "Apache-2.0", "apache license version 2.0": "Apache-2.0",
"bsd": "BSD-3-Clause", "bsd license": "BSD-3-Clause", "new bsd": "BSD-3-Clause", "bsd-3": "BSD-3-Clause",
"3-clause bsd": "BSD-3-Clause", "bsd 3-clause": "BSD-3-Clause", "simplified bsd": "BSD-2-Clause",
"bsd-2": "BSD-2-Clause", "2-clause bsd": "BSD-2-Clause",
"python software foundation license": "PSF-2.0", "psf": "PSF-2.0", "psfl": "PSF-2.0",
"mozilla public license 2.0 (mpl 2.0)": "MPL-2.0", "mpl 2.0": "MPL-2.0", "mpl-2": "MPL-2.0", "mpl2": "MPL-2.0",
"gplv2": "GPL-2.0-only", "gpl-2.0": "GPL-2.0-only", "gpl v2": "GPL-2.0-only", "gpl2": "GPL-2.0-only",
"gplv2+": "GPL-2.0-or-later", "gpl-2.0+": "GPL-2.0-or-later",
"gplv3": "GPL-3.0-only", "gpl-3.0": "GPL-3.0-only", "gpl v3": "GPL-3.0-only", "gpl3": "GPL-3.0-only",
"gplv3+": "GPL-3.0-or-later", "gpl-3.0+": "GPL-3.0-or-later", "gpl": "GPL-3.0-or-later",
"gnu general public license v2 (gplv2)": "GPL-2.0-only", "gnu general public license v3 (gplv3)": "GPL-3.0-only",
"gnu general public license v2 or later (gplv2+)": "GPL-2.0-or-later",
"gnu general public license v3 or later (gplv3+)": "GPL-3.0-or-later",
"lgpl": "LGPL-2.1-or-later", "lgplv2": "LGPL-2.1-only", "lgplv2+": "LGPL-2.1-or-later", "lgpl-2.1": "LGPL-2.1-only",
"lgplv3": "LGPL-3.0-only", "lgplv3+": "LGPL-3.0-or-later", "lgpl-3.0": "LGPL-3.0-only",
"gnu lesser general public license v2 (lgplv2)": "LGPL-2.1-only",
"gnu lesser general public license v3 (lgplv3)": "LGPL-3.0-only",
"gnu lesser general public license v2 or later (lgplv2+)": "LGPL-2.1-or-later",
"agpl": "AGPL-3.0-only", "agplv3": "AGPL-3.0-only", "agpl-3.0": "AGPL-3.0-only",
"gnu affero general public license v3": "AGPL-3.0-only", "gnu affero general public license v3 (agplv3)": "AGPL-3.0-only",
"eclipse public license 2.0": "EPL-2.0", "epl 2.0": "EPL-2.0", "epl-2": "EPL-2.0",
"public domain": "Unlicense", "unlicense": "Unlicense", "the unlicense": "Unlicense", "cc0": "CC0-1.0",
"boost software license 1.0": "BSL-1.0", "zlib license": "Zlib", "zlib": "Zlib",
"business source license 1.1": "BUSL-1.1", "elastic license 2.0": "Elastic-2.0",
"server side public license": "SSPL-1.0", "sspl": "SSPL-1.0",
}
CATEGORY = { # SPDX id prefix -> category
"MIT": "permissive", "MIT-0": "public-domain", "ISC": "permissive", "BSD-2-Clause": "permissive",
"BSD-3-Clause": "permissive", "Apache-2.0": "permissive", "PSF-2.0": "permissive", "Python-2.0": "permissive",
"Zlib": "permissive", "BSL-1.0": "permissive", "X11": "permissive", "BlueOak-1.0.0": "permissive",
"CC-BY-4.0": "permissive", "CC-BY-3.0": "permissive", "Unicode-DFS-2016": "permissive", "Unicode-3.0": "permissive",
"0BSD": "public-domain", "Unlicense": "public-domain", "CC0-1.0": "public-domain", "WTFPL": "public-domain",
"MPL-2.0": "weak-copyleft", "EPL-1.0": "weak-copyleft", "EPL-2.0": "weak-copyleft", "CDDL-1.0": "weak-copyleft",
"CDDL-1.1": "weak-copyleft", "LGPL-2.1": "weak-copyleft", "LGPL-3.0": "weak-copyleft", "CC-BY-SA-4.0": "weak-copyleft",
"GPL-2.0": "strong-copyleft", "GPL-3.0": "strong-copyleft",
"AGPL-3.0": "network-copyleft", "SSPL-1.0": "network-copyleft",
"BUSL-1.1": "source-available", "Elastic-2.0": "source-available", "Commons-Clause": "source-available",
"CC-BY-NC-4.0": "source-available", "CC-BY-NC-SA-4.0": "source-available", "PolyForm-Noncommercial-1.0.0": "source-available",
}
RANK = {"public-domain": 0, "permissive": 1, "weak-copyleft": 2, "strong-copyleft": 3,
"network-copyleft": 4, "source-available": 5, "unknown": 6}
SEV_ORDER = {"HIGH": 0, "WARN": 1, "INFO": 2, "OK": 3}
NOTICE_LICENSES = {"Apache-2.0"}
def unwrap(text):
"""Remove parentheses that wrap the whole expression: '(MIT OR X)' -> 'MIT OR X'."""
t = text.strip()
while t.startswith("(") and t.endswith(")"):
depth = 0
for i, ch in enumerate(t):
depth += ch == "("
depth -= ch == ")"
if depth == 0 and i < len(t) - 1:
return t
t = t[1:-1].strip()
return t
def normalize(token):
t = unwrap(token)
if not t:
return ""
key = t.lower()
if key in ALIASES:
return ALIASES[key]
for spdx in CATEGORY:
if key == spdx.lower() or key.startswith(spdx.lower() + "-") or key == spdx.lower() + "+":
return t if key.startswith(spdx.lower() + "-") else spdx
return t
def category_of(spdx):
if spdx.upper().startswith("LICENSEREF") or spdx.upper().startswith("SEE LICENSE"):
return "unknown"
for prefix, cat in sorted(CATEGORY.items(), key=lambda kv: -len(kv[0])):
if spdx == prefix or spdx.startswith(prefix + "-") or spdx.startswith(prefix + "+"):
return cat
return "unknown"
def classify(expr):
"""Return (normalized expression, chosen license, category, exception) for a license expression."""
raw = (expr or "").strip()
if not raw or raw.upper() in ("UNKNOWN", "NONE", "NOASSERTION", "UNLICENSED", "CUSTOM", "N/A"):
return raw or "(missing)", raw or "(missing)", "unknown", None
if raw.upper().startswith("SEE LICENSE"):
return raw, raw, "unknown", None
whole = unwrap(raw)
if whole.lower() in ALIASES: # long names such as "... v2 or later (LGPLv2+)" are one license
spdx = ALIASES[whole.lower()]
return spdx, spdx, category_of(spdx), None
text = re.sub(r"\s+", " ", whole.replace("/", " OR ").replace(";", " OR "))
exception = None
m = re.search(r"\bWITH\s+([\w.+-]+)", text, re.I)
if m:
exception = m.group(1)
text = text[:m.start()] + text[m.end():]
if re.search(r"\sor\s", text, re.I) and not re.search(r"\sand\s", text, re.I):
parts, mode = re.split(r"\s+or\s+", text, flags=re.I), "or"
elif re.search(r"\sand\s", text, re.I) and not re.search(r"\sor\s", text, re.I):
parts, mode = re.split(r"\s+and\s+", text, flags=re.I), "and"
elif re.search(r"\s(or|and)\s", text, re.I):
parts, mode = re.split(r"\s+(?:or|and)\s+", text, flags=re.I), "and" # mixed: be conservative
else:
parts, mode = [text], "single"
lics = [normalize(p) for p in parts if normalize(p)]
if not lics:
return raw, raw, "unknown", exception
scored = sorted(lics, key=lambda l: RANK[category_of(l)])
chosen = scored[0] if mode == "or" else scored[-1]
cat = category_of(chosen)
if exception and cat == "strong-copyleft" and re.search(r"classpath|linking|gcc|runtime|autoconf|bison|bootloader", exception, re.I):
cat = "weak-copyleft"
joiner = " OR " if mode == "or" else " AND "
return joiner.join(lics) + (f" WITH {exception}" if exception else ""), chosen, cat, exception
def assess(cat, chosen, distribution, project):
proj_cat = "proprietary" if project.lower() == "proprietary" else category_of(normalize(project))
project_spdx = normalize(project)
if cat == "unknown":
return "HIGH", "License is missing or not recognized; find the real license text before shipping."
if cat == "source-available":
return "HIGH", "Source-available or non-commercial terms usually restrict commercial use or hosting; needs a license review."
if project_spdx.startswith("GPL-2.0-only") and chosen.startswith("Apache-2.0"):
return "HIGH", "Apache-2.0 is generally considered incompatible with GPL-2.0-only projects."
gpl3_family = re.match(r"(A|L)?GPL-3\.0", chosen)
if project_spdx.startswith("GPL-2.0-only") and gpl3_family:
return "HIGH", f"{chosen} cannot be combined with a GPL-2.0-only project; the project would need GPL-3.0 terms."
if re.match(r"GPL-3\.0|AGPL-3\.0", project_spdx) and chosen.startswith("GPL-2.0-only"):
return "HIGH", "GPL-2.0-only code cannot be relicensed under GPL-3.0; ask upstream for 'or later' terms or replace it."
if proj_cat in ("strong-copyleft", "network-copyleft") and cat in ("strong-copyleft", "network-copyleft"):
if cat == "network-copyleft" and proj_cat != "network-copyleft" and distribution == "saas":
return "WARN", "AGPL/SSPL in a GPL project run as a service: the whole service source must be offered to users."
return "OK", "Copyleft dependency in a copyleft project; check version compatibility (for example GPL-2.0-only vs GPL-3.0)."
if cat == "network-copyleft":
if distribution == "internal":
return "WARN", "Network copyleft: fine for purely internal use, but exposing it to outside users triggers source obligations."
return "HIGH", "Network copyleft (AGPL/SSPL): offering the service or shipping it requires releasing the corresponding source."
if cat == "strong-copyleft":
if distribution == "distributed":
return "HIGH", "GPL code distributed with your product requires the combined work to be released under the GPL."
if distribution == "saas":
return "WARN", "GPL is not triggered by network use alone, but any future distribution (apps, on-prem, SDKs) would be."
return "INFO", "GPL for internal use only has no distribution obligations; keep it out of anything you ship."
if cat == "weak-copyleft":
if distribution == "distributed":
return "WARN", "Weak copyleft: ship the license, keep the library replaceable (dynamic linking for LGPL), and publish changes to its files."
return "INFO", "Weak copyleft: publish any changes you make to this library's own files if you distribute it later."
if chosen in NOTICE_LICENSES and distribution == "distributed":
return "INFO", "Apache-2.0: include the license and any NOTICE file in your distribution."
return "OK", "Permissive: keep the copyright notice and license text in your third-party notices."
def load(text):
stripped = text.lstrip()
if not stripped:
raise ValueError("input is empty")
if stripped[0] in "[{":
data = json.loads(text)
rows = []
if isinstance(data, dict): # license-checker
for key, val in data.items():
name, _, ver = key.rpartition("@")
if not name:
name, ver = key, ""
lic = val.get("licenses", "") if isinstance(val, dict) else val
if isinstance(lic, list):
lic = " OR ".join(lic)
rows.append({"name": name, "version": ver, "license": str(lic), "scope": "runtime"})
return rows
for item in data:
if not isinstance(item, dict):
raise ValueError("JSON list items must be objects")
low = {k.lower(): v for k, v in item.items()}
lic = low.get("license", low.get("licenses", ""))
if isinstance(lic, list):
lic = " OR ".join(lic)
rows.append({"name": str(low.get("name", "")), "version": str(low.get("version", "")),
"license": str(lic or ""), "scope": str(low.get("scope", "runtime") or "runtime").lower()})
return rows
reader = csv.DictReader(io.StringIO(text))
fields = [f.strip().lower() for f in (reader.fieldnames or [])]
if "name" not in fields or "license" not in fields:
raise ValueError(f"CSV needs a header with at least 'name' and 'license' columns, got: {', '.join(fields) or '(none)'}")
rows = []
for r in reader:
r = {(k or "").strip().lower(): (v or "").strip() for k, v in r.items()}
if not r.get("name"):
continue
rows.append({"name": r["name"], "version": r.get("version", ""), "license": r.get("license", ""),
"scope": (r.get("scope") or "runtime").lower()})
return rows
def audit(rows, args):
allow = {normalize(x) for x in args.allow.split(",") if x.strip()} if args.allow else set()
deny = {normalize(x) for x in args.deny.split(",") if x.strip()} if args.deny else set()
results, seen = [], {}
for r in rows:
expr, chosen, cat, exc = classify(r["license"])
sev, why = assess(cat, chosen, args.distribution, args.project_license)
if exc:
why += f" License exception {exc} applies; read its exact scope."
if chosen in deny:
sev, why = "HIGH", f"{chosen} is on your deny list."
elif chosen in allow and sev != "OK":
sev, why = "OK", f"{chosen} is on your allow list ({why})"
dev = r["scope"] in ("dev", "development", "test", "build")
if dev and sev in ("HIGH", "WARN") and chosen not in deny:
sev, why = "INFO", "Dev-only (not shipped): " + why
results.append({"name": r["name"], "version": r["version"], "license_raw": r["license"],
"license": expr, "effective": chosen, "category": cat, "scope": "dev" if dev else "runtime",
"severity": sev, "reason": why})
seen.setdefault(r["name"].lower(), set()).add(expr)
notes = []
for name, lics in seen.items():
if len(lics) > 1:
notes.append(f"{name}: different licenses across versions ({', '.join(sorted(lics))}); check which version you ship.")
results.sort(key=lambda x: (SEV_ORDER[x["severity"]], -RANK[x["category"]], x["name"].lower()))
counts = {s: sum(1 for x in results if x["severity"] == s) for s in SEV_ORDER}
cats = {}
for x in results:
cats[x["category"]] = cats.get(x["category"], 0) + 1
attribution = sorted({x["effective"] for x in results if x["scope"] == "runtime" and x["category"] in ("permissive", "weak-copyleft")})
return {"project_license": args.project_license, "distribution": args.distribution, "packages": len(results),
"severity_counts": counts, "category_counts": cats, "attribution_licenses": attribution,
"notes": notes, "results": results}
def print_text(rep):
print(f"Dependency license audit: {rep['packages']} packages, project license {rep['project_license']}, distribution {rep['distribution']}")
c = rep["severity_counts"]
print(f"HIGH {c['HIGH']} WARN {c['WARN']} INFO {c['INFO']} OK {c['OK']}")
print("Categories: " + ", ".join(f"{k} {v}" for k, v in sorted(rep["category_counts"].items(), key=lambda kv: RANK[kv[0]])))
print()
w = max([len(f"{x['name']}@{x['version']}") for x in rep["results"]] + [7])
print(f"{'SEV':<5} {'PACKAGE':<{w}} {'LICENSE':<28} {'CATEGORY':<17} SCOPE")
for x in rep["results"]:
pkg = f"{x['name']}@{x['version']}" if x["version"] else x["name"]
print(f"{x['severity']:<5} {pkg:<{w}} {x['license'][:28]:<28} {x['category']:<17} {x['scope']}")
findings = [x for x in rep["results"] if x["severity"] in ("HIGH", "WARN", "INFO")]
if findings:
print("\nFindings:")
for x in findings:
print(f"- [{x['severity']}] {x['name']} ({x['license_raw'] or 'no license field'}): {x['reason']}")
if rep["notes"]:
print("\nNotes:")
for n in rep["notes"]:
print(f"- {n}")
if rep["attribution_licenses"]:
print("\nThird-party notices must include license texts for: " + ", ".join(rep["attribution_licenses"]))
print("\nTriage aid only, not legal advice.")
def main():
p = argparse.ArgumentParser(description="Audit dependency licenses.", add_help=True)
p.add_argument("path", help="CSV or JSON file, or - for stdin")
p.add_argument("--project-license", default="proprietary")
p.add_argument("--distribution", default="saas", choices=["saas", "distributed", "internal"])
p.add_argument("--allow", default="")
p.add_argument("--deny", default="")
p.add_argument("--json", action="store_true")
args = p.parse_args()
try:
text = sys.stdin.read() if args.path == "-" else open(args.path, encoding="utf-8").read()
rows = load(text)
except FileNotFoundError:
print(f"error: file not found: {args.path}", file=sys.stderr)
return 2
except (ValueError, json.JSONDecodeError) as e:
print(f"error: {e}", file=sys.stderr)
return 2
if not rows:
print("error: no dependencies found in the input", file=sys.stderr)
return 2
rep = audit(rows, args)
if args.json:
print(json.dumps(rep, indent=2))
else:
print_text(rep)
return 1 if rep["severity_counts"]["HIGH"] else 0
if __name__ == "__main__":
sys.exit(main())Checks READMEs and Markdown docs for accessibility: missing or file-name alt text, heading levels, vague or duplicate link text, bare URLs, empty table headers, color-only instructions, ALL CAPS, and emoji bullets, then writes purposeful alt text and a corrected file. Includes a tested stdlib Python checker for CI.
---
name: markdown-accessibility-checker
description: Reviews Markdown documentation (READMEs, docs sites, wikis, blog posts, changelogs) for accessibility - missing or poor image alt text, heading structure, vague or duplicate link text, bare URLs, empty table headers, color-only instructions, ALL CAPS, and emoji bullets - then writes good alt text for each image from context and returns a fix list or a corrected file. Use when a user asks to "check my README for accessibility", "write alt text for these images", "make our docs screen reader friendly", or wants an accessibility check for Markdown in CI.
---
# Markdown Accessibility Checker
You make Markdown documents work for everyone, including people who use screen readers, keyboard navigation, magnification, or reading aids. You find the problems a script can detect, then do the part a script cannot: write meaningful alt text and link text that fit the document's purpose.
## Files in this skill
- `scripts/check_md_a11y.py` - scans Markdown files or folders and reports problems with line numbers and severities (Python 3 standard library only)
- `references/alt-text-guide.md` - how to write alt text by image type (screenshots, diagrams, charts, photos, logos, badges, decorative images)
- `references/structure-and-links.md` - headings, link text, tables, lists, emphasis, and what each rule in the script means
- `templates/a11y-review.md` - the review report and the alt text worksheet
- `examples/example-readme-review.md` - a worked review of a project README
## Workflow
### 1. Scope the review
Ask which files or folders to check, who reads them (developers, end users, students), and whether the user wants a report, a corrected file, or both. If the documents are rendered by a site generator, ask whether it adds its own page title (then the Markdown may not need an H1).
### 2. Run the checker
```bash
python3 scripts/check_md_a11y.py README.md docs/
python3 scripts/check_md_a11y.py docs/ --json
python3 scripts/check_md_a11y.py docs/ --fail-on warn # for CI
```
It ignores fenced code blocks and inline code. HIGH means content is lost for some readers (missing alt, empty link, empty heading), WARN should be fixed, INFO needs a human look. If you cannot run the script, check the same rules by reading the file and say so.
### 3. Write alt text
For every image flagged, and for every image with existing alt text, decide its purpose using `references/alt-text-guide.md`:
- Read the text around it: what is the image for at this point in the document?
- If you can see the image, describe what matters for that purpose. If you cannot, ask the user, or write a clearly marked draft from the file name and context (for example "DRAFT: check against the image").
- Decorative images get empty alt text; in Markdown prefer removing them, or use HTML `<img alt="">`.
Fill in the alt text worksheet in `templates/a11y-review.md`.
### 4. Fix structure and links
Apply `references/structure-and-links.md`: one H1, no skipped levels, real headings instead of bold lines, descriptive unique link text, tables with headers, instructions that do not rely on color alone.
### 5. Deliver
- Report: fill in `templates/a11y-review.md`, as in `examples/example-readme-review.md`, with the most important fixes first.
- Corrected file: return the full Markdown with fixes applied and a short change list. Do not change meaning, code samples, or links' destinations.
- Rerun the checker on the corrected file and include the before and after counts.
## Rules
- Never invent what an image shows. If you have not seen it, mark the alt text as a draft.
- Keep alt text short (usually under 150 characters); put long explanations (charts, diagrams) in the text or a caption and say so in the alt.
- Do not start alt text with "image of" or "picture of".
- Remember that automated checks find only part of the problems; always mention what still needs a manual check (color contrast in images, reading order, video captions).
- Respect the author's voice; fix accessibility, not style.
FILE:references/alt-text-guide.md
# Alt text guide
Alt text replaces the image for people who cannot see it. Ask: "If I read this document aloud, what would I say at this point instead of showing the image?" The answer depends on why the image is there, not only on what is in it.
## Basics
- Be specific and short: usually 5 to 20 words, rarely more than 150 characters.
- Lead with the important part. Do not start with "image of", "picture of", or "screenshot of" unless the type itself matters ("Screenshot" can help when the reader needs to find the same screen).
- End with a period so screen readers pause.
- Do not repeat the caption or the sentence next to the image word for word.
- Include text that appears in the image if the reader needs it (button labels, error messages).
## By image type
### Screenshots of an app or website
State the screen and the one thing the reader should notice.
- Weak: "screenshot"
- Better: "Settings screen with the Units toggle set to Metric."
### Hero images and banners
If it only sets the mood, it is decorative. If it shows the product, describe it in one sentence.
- "Trailhead Weather on a phone, showing a three-day forecast for a mountain pass."
### Diagrams and flowcharts
Give the main idea in the alt, and the full structure in the text or a list right after the image.
- Alt: "Architecture diagram: the mobile app calls the API, which reads from a forecast cache and a weather provider. Details in the list below."
### Charts and graphs
Give the type, the subject, and the key takeaway; put the data in a table if readers need numbers.
- "Line chart of daily active users, January to June 2026, rising from 2,000 to 9,500 with a dip in April."
### Photos
Describe what matters for the document's purpose: subject, action, setting.
- "Snow-covered ridge trail with orange marker poles leading to a hut."
### Logos
Use the organization or product name, plus "logo" only if it helps: "Trailhead Weather logo."
When the logo is a link, describe the destination: "Trailhead Weather home."
### Badges (build status, version, license)
Write the information the badge shows: "Build: passing", "License: MIT", "Version 2.4.1".
### Icons inside text or links
If the icon has a function, describe the function ("Download"); if it repeats nearby text, it is decorative.
### Animated GIFs and short videos
Describe what happens: "Animation: dragging a pin onto the map adds a new trailhead to the list." Videos also need captions and, for important content, a text alternative.
### Decorative images
Dividers, background textures, and purely visual flourishes. In HTML use `alt=""` so screen readers skip them. In Markdown, `` is ambiguous (often a mistake), so prefer removing the image or using `<img src="divider.png" alt="">`.
## Common mistakes
- File names as alt text: "IMG_2041.jpg", "hero-final-v3".
- One word: "image", "logo", "chart".
- Keyword stuffing for search engines.
- Different alt text for the same image used in several places.
- Describing colors or layout when the reader needs the information, not the look.
FILE:references/structure-and-links.md
# Structure, links, and what each rule means
Screen reader users often jump through a document by headings or by links. Good structure is what makes that possible.
## Headings
| Rule | Severity | Why | Fix |
|---|---|---|---|
| empty-heading | HIGH | An empty heading is announced as "heading level 2" with nothing after it. | Add text or remove the #. |
| no-h1 | WARN | The page has no title in the heading outline. | Start with one `# Title`. Skip this if the site generator adds the H1 from front matter. |
| multiple-h1 | WARN | Several titles confuse the outline. | Keep one H1; use `##` for sections. |
| heading-skip | WARN | Jumping from H2 to H4 suggests a missing section. | Use the next level down (H3). Pick levels for structure, not for font size. |
| fake-heading | INFO | A bold line looks like a heading but is not in the outline. | Use a real heading if it starts a section; keep bold if it is just emphasis. |
| no-headings | INFO | Long text without headings is hard to scan. | Add section headings every few paragraphs. |
## Links
| Rule | Severity | Why | Fix |
|---|---|---|---|
| empty-link | HIGH | The link has no accessible name. | Add descriptive text. |
| vague-link | WARN | "Click here" or "read more" means nothing in a list of links. | Name the destination: "download the latest release", "setup guide". |
| same-text-different-url | WARN | Two links called "docs" that go to different places are indistinguishable. | Make each text unique: "API reference", "setup guide". |
| url-as-link-text | INFO | Long URLs are read character by character. | Use a short description; keep the URL only if readers must copy it. |
| bare-url | INFO | Same as above, and not always clickable. | Wrap in a descriptive link. |
Link text tips: put the important words first, keep it short, say when a link opens a file ("pricing sheet (PDF, 2 MB)"), and do not use "link" in the text.
## Images
| Rule | Severity | Why |
|---|---|---|
| img-missing-alt | HIGH | `` gives no alternative; some readers announce the file name. |
| img-no-alt | HIGH | `<img>` without an alt attribute; screen readers fall back to the file name. |
| img-generic-alt | WARN | "image", "screenshot", "logo" alone carry no information. |
| img-filename-alt | WARN | "IMG_2041.jpg" is not a description. |
| img-long-alt | WARN | Very long alt text cannot be navigated; move details to the text. |
| img-redundant-alt | INFO | "Image of" is redundant; screen readers already say "image". |
| img-decorative | INFO | `alt=""` hides the image; confirm it is truly decorative. |
See `alt-text-guide.md` for how to write the text.
## Tables
- table-empty-header (WARN): every column needs a header so screen readers can announce "Pro, Forecast days, 10". Add a header such as "Feature" to the first column.
- Use tables for data, not for layout. Keep them small; split very wide tables.
## Other rules
- color-only (WARN): "Press the green button" fails for color-blind readers and screen reader users. Name the button: "Select Save (the green button)".
- all-caps (INFO): long ALL CAPS text is harder to read, and some screen readers spell short caps words letter by letter. Use normal case and bold for emphasis.
- emoji-bullet (INFO): emoji are read aloud by name ("white heavy check mark Hourly forecasts"). Use normal list markers, and keep emoji rare and meaningful.
## Not checked by the script (check by hand)
- Whether alt text is accurate and useful.
- Color contrast and text inside images (screenshots with tiny text, charts with low contrast).
- Reading order of complex layouts, and HTML blocks embedded in Markdown.
- Captions and transcripts for videos and audio.
- Plain language: short sentences, defined jargon, and descriptive section names.
FILE:templates/a11y-review.md
# Accessibility Review: <document or folder>
Date: <YYYY-MM-DD>
Files checked: <n> (<list or folder>)
Audience: <who reads these docs>
Checker run: `python3 scripts/check_md_a11y.py <paths>` -> HIGH <n>, WARN <n>, INFO <n>
## Top fixes (most impact first)
1. <fix> (<file:line>)
2. <fix>
3. <fix>
## Alt text worksheet
| File:line | Image | Purpose in the document | Current alt | New alt | Seen the image? |
|---|---|---|---|---|---|
| | | informative / functional / decorative / complex | | | yes / draft |
## Headings
- Current outline: <H1 ... / H2 ... / H4 ...>
- Proposed outline: <H1 ... / H2 ... / H3 ...>
## Links
| File:line | Current text | New text | Destination |
|---|---|---|---|
| | | | |
## Other findings
| File:line | Rule | Change |
|---|---|---|
| | | |
## After fixes
Checker rerun: HIGH <n>, WARN <n>, INFO <n> (remaining INFO items explained below).
- <INFO item: reason it is acceptable>
## Manual checks still needed
- <for example: contrast of text inside screenshots, video captions>
FILE:examples/example-readme-review.md
# Example: accessibility review of a project README
## The request
"Our open source app's README gets a lot of traffic. A user who relies on a screen reader said it is hard to follow. Can you check it and fix it? The hero image is the app on a phone with the 3-day forecast; app-main.png is the main screen with hourly temperature, wind, and snow line; the badge is our build status."
## Checker run (before)
```
$ python3 scripts/check_md_a11y.py README.md
Markdown accessibility check: 1 file(s), 3 image(s)
HIGH 2 WARN 8 INFO 5
README.md (3 images, 4 headings): 15 finding(s)
3 HIGH img-missing-alt Image hero.png has empty alt text; describe it, or confirm it is purely decorative.
5 INFO all-caps Long ALL CAPS text is hard to read and some screen readers spell it letter by letter.
9 WARN vague-link Link text "here" does not say where it goes; describe the destination.
9 WARN vague-link Link text "click here" does not say where it goes; describe the destination.
11 WARN img-generic-alt Alt text "screenshot" for app-main.png is generic; say what the image shows or why it is there.
13 HIGH img-no-alt <img src="docs/img/badge.svg"> has no alt attribute.
15 WARN heading-skip Heading level jumps from H2 to H4 ("Configuration"); use H3.
17 WARN color-only Instruction relies on color; add the label, shape, or position too.
19 INFO fake-heading Bold line "Features" looks like a heading; use a real # heading if it starts a section.
21 INFO emoji-bullet Emoji used as a bullet is read aloud by name; use a list marker and put meaning in words.
22 INFO emoji-bullet Emoji used as a bullet is read aloud by name; use a list marker and put meaning in words.
24 WARN table-empty-header Table header has 1 empty cell(s); every column needs a header.
28 WARN same-text-different-url Link text "docs" points to 2 different URLs; make each text unique.
28 INFO bare-url Bare URL https://example.org/docs; screen readers spell it out. Use [descriptive text](https://example.org/docs).
35 WARN multiple-h1 2 H1 headings (lines 1, 35); keep one title and use ## for sections.
```
The `` inside the bash code block and the inline code example on the last line were correctly ignored.
## The review (filled template, shortened)
# Accessibility Review: README.md
Audience: hikers and contributors, many arriving from search.
Checker run: HIGH 2, WARN 8, INFO 5.
### Top fixes
1. Give the hero image and the build badge real alt text (lines 3 and 13).
2. Replace "here" and "click here" with link text that names the destination (line 9).
3. Fix the outline: "Configuration" becomes H3, "Features" becomes a real H3, and the second H1 "Contributing" becomes H2.
4. Name the Save button instead of only its color (line 17).
5. Give the first table column a header (line 24).
### Alt text worksheet
| File:line | Image | Purpose | Current alt | New alt | Seen? |
|---|---|---|---|---|---|
| README.md:3 | hero.png | informative: shows the product | (empty) | Trailhead Weather on a phone, showing a three-day forecast for a mountain pass. | described by the user |
| README.md:11 | app-main.png | informative: what the main screen shows | screenshot | Main screen with hourly temperature, wind, and snow line for the selected trailhead. | described by the user |
| README.md:13 | badge.svg | informative: build status | (no alt) | Build: passing | described by the user |
### Headings
- Before: H1 Trailhead Weather / H2 Install / H4 Configuration / H1 Contributing (plus a bold "Features" line).
- After: H1 Trailhead Weather / H2 Install / H3 Configuration / H3 Features / H2 Contributing.
### Links
| Line | Current text | New text | Destination |
|---|---|---|---|
| 9 | here | Download the latest release | https://example.org/releases |
| 9 | click here | setup guide | docs/guides/setup.md |
| 28 | docs (bare URL + "docs") | API reference | https://example.org/docs |
| 28 | docs | setup guide | docs/guides/setup.md |
### Other findings
| Line | Rule | Change |
|---|---|---|
| 5 | all-caps | Normal case inside the bold note. |
| 17 | color-only | "Select Save (the green button at the bottom of Settings)". |
| 21-22 | emoji-bullet | Normal "-" list items. |
| 24 | table-empty-header | First column header "Feature". |
### Corrected README (excerpt)
```markdown
# Trailhead Weather
Trailhead Weather shows mountain forecasts for hikers. **Note: this project is still in beta and things will break often.**
## Install
[Download the latest release](https://example.org/releases), or follow the [setup guide](docs/guides/setup.md).
<img src="docs/img/badge.svg" width="90" alt="Build: passing">
### Configuration
Select Save (the green button at the bottom of Settings) to save your settings.
### Features
- Hourly forecasts
- Avalanche bulletins
| Feature | Free | Pro |
|---|---|---|
| Forecast days | 3 | 10 |
```
### After fixes
```
$ python3 scripts/check_md_a11y.py README-fixed.md
Markdown accessibility check: 1 file(s), 3 image(s)
HIGH 0 WARN 0 INFO 0
```
### Manual checks still needed
- The hero image has small text on the phone screen; the alt text carries the key message, so that is fine here.
- If the README is ever turned into a docs site page, confirm the site does not add a second H1.
## Why this is a good answer
- Alt text describes each image's purpose, using what the user said the images show instead of guessing.
- Links and headings are fixed for how screen reader users navigate: by link list and by heading outline.
- The fix is verified by rerunning the checker, and the remaining manual checks are listed honestly.
FILE:scripts/check_md_a11y.py
#!/usr/bin/env python3
"""Check Markdown files for common accessibility problems.
Usage:
python3 check_md_a11y.py README.md docs/ # files and folders (*.md, *.markdown)
python3 check_md_a11y.py docs/ --json
cat page.md | python3 check_md_a11y.py -
python3 check_md_a11y.py docs/ --fail-on warn # exit 1 on WARN too (for CI)
Checks (code blocks and inline code are ignored):
images missing alt text, file-name or generic alt ("image", "screenshot"),
redundant "image of ..." openers, very long alt, HTML <img> without alt,
empty alt (fine only for decorative images)
headings no H1 or more than one H1, skipped levels (H2 -> H4), empty headings,
bold lines used as fake headings
links empty link text, vague text ("click here", "read more", "here"),
bare URLs, the same text pointing to different URLs
tables empty header cells
other ALL CAPS sentences, emoji used as bullets, color-only instructions
Severities: HIGH (blocks understanding), WARN (should fix), INFO (check by hand).
Exit code: 0 no HIGH, 1 HIGH found (or WARN with --fail-on warn), 2 usage error.
Standard library only.
"""
import argparse
import json
import os
import re
import sys
GENERIC_ALT = {"image", "img", "picture", "pic", "photo", "screenshot", "screen shot", "graphic", "icon",
"logo", "diagram", "chart", "figure", "untitled", "alt", "alt text", "banner", "placeholder"}
VAGUE_LINK = {"click here", "here", "click", "read more", "more", "learn more", "link", "this", "this link",
"this page", "go", "details", "more info", "info", "see here", "check it out"}
REDUNDANT_START = re.compile(r"^(an? )?(image|picture|photo|screenshot|graphic) (of|showing)\b", re.I)
FILE_ALT = re.compile(r"^[\w\-. ]+\.(png|jpe?g|gif|svg|webp|bmp|tiff?)$|^(img|image|dsc|screenshot)[_\- ]?\d+", re.I)
IMG_MD = re.compile(r"!\[([^\]]*)\]\(\s*<?([^)\s>]*)>?(?:\s+\"[^\"]*\")?\s*\)")
IMG_REF = re.compile(r"!\[([^\]]*)\]\[[^\]]*\]")
LINK_MD = re.compile(r"(?<!!)\[([^\]]*)\]\(\s*<?([^)\s>]*)>?(?:\s+\"[^\"]*\")?\s*\)")
HTML_IMG = re.compile(r"<img\b[^>]*>", re.I)
BARE_URL = re.compile(r"(?<![(<\"'=\[])\bhttps?://[^\s)>\]]+")
COLOR_ONLY = re.compile(r"\b(click|press|tap|select|choose|use)\b[^.]{0,30}\b(red|green|blue|yellow|orange|purple)\b(?! *\()", re.I)
EMOJI_BULLET = re.compile(r"^\s*[\u2705\u274c\u2b50\u27a1\u2714\u2716\U0001F300-\U0001FAFF]\ufe0f?\s+\S")
def strip_inline_code(line):
return re.sub(r"`+[^`]*`+", lambda m: " " * len(m.group(0)), line)
def check_text(text, name):
findings = []
def add(sev, line, rule, msg):
findings.append({"file": name, "line": line, "severity": sev, "rule": rule, "message": msg})
lines = text.replace("\r\n", "\n").split("\n")
in_fence, fence = False, ""
headings, link_targets = [], {}
images = 0
prev_blank = True
for no, raw in enumerate(lines, 1):
stripped = raw.strip()
m = re.match(r"^\s{0,3}(`{3,}|~{3,})", raw)
if m:
if not in_fence:
in_fence, fence = True, m.group(1)[0] * 3
elif stripped.startswith(fence):
in_fence = False
continue
if in_fence or raw.startswith(" ") and prev_blank and not stripped.startswith(("-", "*", "+")):
prev_blank = not stripped if not in_fence else prev_blank
continue
line = strip_inline_code(raw)
hm = re.match(r"^\s{0,3}(#{1,6})(\s+|$)(.*?)\s*#*\s*$", line)
if hm:
level, title = len(hm.group(1)), hm.group(3).strip()
headings.append((no, level, title))
if not title:
add("HIGH", no, "empty-heading", f"Empty H{level} heading; screen reader users navigate by headings.")
elif re.fullmatch(r"\s*(\*\*|__)[^*_]{2,80}(\*\*|__):?\s*", line) and prev_blank:
add("INFO", no, "fake-heading", f"Bold line \"{stripped.strip('*_:')}\" looks like a heading; use a real # heading if it starts a section.")
images += len(IMG_MD.findall(line)) + len(IMG_REF.findall(line)) + len(HTML_IMG.findall(line))
for alt, src in IMG_MD.findall(line):
check_alt(add, no, alt, src)
for alt in IMG_REF.findall(line):
check_alt(add, no, alt, "(reference)")
for tag in HTML_IMG.findall(line):
am = re.search(r"\balt\s*=\s*(\"([^\"]*)\"|'([^']*)'|(\S+))", tag, re.I)
src = re.search(r"\bsrc\s*=\s*[\"']?([^\"'\s>]+)", tag, re.I)
src = src.group(1) if src else "?"
if not am:
add("HIGH", no, "img-no-alt", f"<img src=\"{src}\"> has no alt attribute.")
else:
check_alt(add, no, am.group(2) if am.group(2) is not None else (am.group(3) if am.group(3) is not None else am.group(4)), src, html=True)
no_images = IMG_MD.sub(" ", line)
for text_, url in LINK_MD.findall(no_images):
label = re.sub(r"[*_`]", "", text_).strip()
if not label:
add("HIGH", no, "empty-link", f"Link to {url or '(no url)'} has no text; screen readers will read the URL or nothing.")
elif label.lower().rstrip(".!:") in VAGUE_LINK:
add("WARN", no, "vague-link", f"Link text \"{label}\" does not say where it goes; describe the destination.")
elif re.fullmatch(r"https?://\S+", label):
add("INFO", no, "url-as-link-text", f"Link text is a raw URL ({label[:60]}); use a short description.")
if label:
link_targets.setdefault(label.lower(), {}).setdefault(url, no)
for url in BARE_URL.findall(LINK_MD.sub(" ", no_images)):
add("INFO", no, "bare-url", f"Bare URL {url[:60]}; screen readers spell it out. Use [descriptive text]({url[:40]}).")
if "|" in line and re.match(r"^\s*\|?\s*:?-{3,}:?\s*(\|\s*:?-{3,}:?\s*)+\|?\s*$", line) and no > 1:
header = lines[no - 2].strip().strip("|")
cells = [c.strip() for c in header.split("|")]
if any(not c for c in cells):
add("WARN", no - 1, "table-empty-header", f"Table header has {sum(1 for c in cells if not c)} empty cell(s); every column needs a header.")
run = longest = 0
for w in re.findall(r"[A-Za-z]+", line):
run = run + 1 if w.isupper() and (len(w) > 1 or w in "AI") else 0
longest = max(longest, run)
if longest >= 5 and not hm:
add("INFO", no, "all-caps", "Long ALL CAPS text is hard to read and some screen readers spell it letter by letter.")
if EMOJI_BULLET.match(line):
add("INFO", no, "emoji-bullet", "Emoji used as a bullet is read aloud by name; use a list marker and put meaning in words.")
cm = COLOR_ONLY.search(line)
if cm:
between = line[cm.end(1):cm.start(2)]
inside_parens = line[:cm.start(2)].count("(") > line[:cm.start(2)].count(")")
names_label = re.match(r"\s+(\*\*|\"|')?[A-Z]", between) # "Select Save ..." or "Press **Save**"
if inside_parens or names_label:
cm = None
if cm:
add("WARN", no, "color-only", "Instruction relies on color; add the label, shape, or position too.")
prev_blank = not stripped
h1 = [h for h in headings if h[1] == 1]
if headings and not h1:
add("WARN", headings[0][0], "no-h1", "Document has no H1; start with one # title.")
if len(h1) > 1:
add("WARN", h1[1][0], "multiple-h1", f"{len(h1)} H1 headings (lines {', '.join(str(h[0]) for h in h1)}); keep one title and use ## for sections.")
for (pno, plev, _), (no, lev, title) in zip(headings, headings[1:]):
if lev > plev + 1:
add("WARN", no, "heading-skip", f"Heading level jumps from H{plev} to H{lev} (\"{title[:50]}\"); use H{plev + 1}.")
if not headings and len([l for l in lines if l.strip()]) > 15:
add("INFO", 1, "no-headings", "Long document with no headings; add headings so readers can navigate.")
for label, targets in link_targets.items():
if len(targets) > 1 and label not in VAGUE_LINK:
first = min(targets.values())
add("WARN", first, "same-text-different-url", f"Link text \"{label}\" points to {len(targets)} different URLs; make each text unique.")
order = {"HIGH": 0, "WARN": 1, "INFO": 2}
findings.sort(key=lambda f: (f["line"], order[f["severity"]]))
return findings, len(headings), images
def check_alt(add, no, alt, src, html=False):
a = alt.strip()
base = os.path.basename(src) if src else ""
if a == "" and html:
add("INFO", no, "img-decorative", f"<img src=\"{base}\"> has alt=\"\" (marked decorative); confirm it carries no information.")
return
if a == "" and src:
add("HIGH", no, "img-missing-alt", f"Image {base or src} has empty alt text; describe it, or confirm it is purely decorative.")
return
low = a.lower().rstrip(".")
if low in GENERIC_ALT:
add("WARN", no, "img-generic-alt", f"Alt text \"{a}\" for {base} is generic; say what the image shows or why it is there.")
elif FILE_ALT.search(a) or (base and low == os.path.splitext(base)[0].lower()):
add("WARN", no, "img-filename-alt", f"Alt text \"{a}\" looks like a file name; describe the image instead.")
elif REDUNDANT_START.match(a):
add("INFO", no, "img-redundant-alt", f"Alt text starts with \"{' '.join(a.split()[:3])}\"; screen readers already say 'image', so start with the content.")
if len(a) > 150:
add("WARN", no, "img-long-alt", f"Alt text is {len(a)} characters; keep it under about 150 and move details to the surrounding text or a caption.")
def collect(paths):
files = []
for p in paths:
if p == "-":
files.append("-")
elif os.path.isdir(p):
for root, dirs, names in os.walk(p):
dirs[:] = sorted(d for d in dirs if not d.startswith(".") and d != "node_modules")
files += [os.path.join(root, n) for n in sorted(names) if n.lower().endswith((".md", ".markdown"))]
elif os.path.isfile(p):
files.append(p)
else:
raise FileNotFoundError(p)
return files
def main():
ap = argparse.ArgumentParser(description="Markdown accessibility checker")
ap.add_argument("paths", nargs="+", help="Markdown files, folders, or - for stdin")
ap.add_argument("--json", action="store_true")
ap.add_argument("--fail-on", choices=["high", "warn"], default="high")
args = ap.parse_args()
try:
files = collect(args.paths)
except FileNotFoundError as e:
print(f"error: path not found: {e}", file=sys.stderr)
return 2
if not files:
print("error: no Markdown files (*.md, *.markdown) found", file=sys.stderr)
return 2
all_findings, stats = [], []
for f in files:
text = sys.stdin.read() if f == "-" else open(f, encoding="utf-8", errors="replace").read()
name = "<stdin>" if f == "-" else f
found, nhead, images = check_text(text, name)
stats.append({"file": name, "images": images, "headings": nhead, "findings": len(found)})
all_findings += found
counts = {s: sum(1 for x in all_findings if x["severity"] == s) for s in ("HIGH", "WARN", "INFO")}
if args.json:
print(json.dumps({"files": stats, "counts": counts, "findings": all_findings}, indent=2))
else:
print(f"Markdown accessibility check: {len(files)} file(s), {sum(s['images'] for s in stats)} image(s)")
print(f"HIGH {counts['HIGH']} WARN {counts['WARN']} INFO {counts['INFO']}")
for s in stats:
items = [x for x in all_findings if x["file"] == s["file"]]
print(f"\n{s['file']} ({s['images']} images, {s['headings']} headings): {len(items)} finding(s)")
for x in items:
print(f" {x['line']:>4} {x['severity']:<4} {x['rule']:<24} {x['message']}")
print("\nAutomated checks cover only part of accessibility; review alt text meaning, reading order, and contrast by hand.")
bad = counts["HIGH"] + (counts["WARN"] if args.fail_on == "warn" else 0)
return 1 if bad else 0
if __name__ == "__main__":
sys.exit(main())Turns a list of meeting topics into a timed agenda with clock times, owners, and an outcome per item, checks it fits the slot with a buffer, flags late or rushed decisions, too many status updates, oversized items, and missing breaks, and writes the agenda to send. Includes a tested stdlib Python timeboxer with a fit mode.
---
name: meeting-agenda-timeboxer
description: Turns a rough list of meeting topics into a timeboxed agenda with start and end times, owners, and a clear outcome for every item (inform, discuss, decide, brainstorm, review), checks that it fits the slot with a buffer, flags agenda smells such as late or rushed decisions, too many status updates, oversized items, and missing breaks, and writes a ready-to-send agenda and pre-read. Use when a user asks "help me plan this meeting", "will this agenda fit in an hour?", "make this agenda tighter", or shares a list of topics for a team meeting, workshop, offsite, or review.
---
# Meeting Agenda Timeboxer
You help people run meetings that end on time and produce decisions. You turn a list of topics into a realistic schedule, give every item an owner and a purpose, move status updates out of the room, and make the time math visible so nobody discovers at minute 55 that the main decision has no time left.
## Files in this skill
- `scripts/timebox_agenda.py` - schedules every item, checks the fit against the slot and buffer, flags agenda smells, and can fit items to the slot in 5-minute steps (Python 3 standard library only)
- `references/timeboxing-principles.md` - how long things really take, item outcomes, ordering, buffers, and breaks
- `references/facilitation-moves.md` - techniques to keep each kind of item inside its timebox, and what to do when one runs over
- `templates/agenda.md` - the agenda input format and the agenda to send
- `examples/example-roadmap-meeting.md` - a worked fix of an overloaded one-hour planning meeting
## Workflow
### 1. Collect the facts
Ask, or assume and say so:
- Purpose of the meeting in one sentence: what must be true when it ends?
- Start time and length (or end time), number of attendees, in person or remote.
- Topics, each with an owner and what the group must do with it: inform, update, discuss, brainstorm, review, or decide.
- Hard constraints: people who join late or leave early, fixed breaks.
### 2. Write the draft agenda
Use the input format in `templates/agenda.md` (one line per item: `minutes | topic | owner | outcome`). Use `?` for a duration nobody knows yet. Estimate durations with `references/timeboxing-principles.md`.
### 3. Run the timeboxer
```bash
python3 scripts/timebox_agenda.py agenda.txt
python3 scripts/timebox_agenda.py agenda.txt --fit # scale to the slot in 5-minute steps
python3 scripts/timebox_agenda.py agenda.json --json
```
It prints the schedule with clock times, the buffer, and findings: HIGH when the agenda does not fit, WARN for smells, INFO for unplanned time and the meeting cost in person-hours. `--fit` trims status items first, then scales everything else; it never reorders items, so check the order yourself afterwards.
If you cannot run the script, do the clock math by hand, item by item, and say so.
### 4. Fix the agenda, not just the numbers
Work through the findings in this order:
1. Remove or shrink status updates and announcements: move them into a written pre-read.
2. Put decisions early (first half), and give each one at least 10 minutes plus a written proposal sent beforehand.
3. Give every item an owner and an outcome; drop "AOB" or turn it into a 5-minute parking lot review.
4. Split items over 45 minutes; add a break every 60 to 90 minutes.
5. If it still does not fit, propose what to cut or move to another meeting or to async, and say what is lost.
Rerun the script until there are no HIGH findings and every WARN is either fixed or explained.
### 5. Deliver
Fill in the agenda to send from `templates/agenda.md`: purpose, desired outcomes, the timed schedule, pre-reads with owners and deadlines, and the parking lot rule. Add one or two facilitation tips per risky item from `references/facilitation-moves.md`. Follow the style of `examples/example-roadmap-meeting.md`.
## Rules
- Never make the math work by shrinking a decision below what it needs; cut or move something else, or recommend a longer slot.
- Keep a buffer: about 5 minutes per hour, 10 percent for workshops.
- Say plainly when a meeting should not happen (no decisions, nothing to discuss) and suggest the async alternative.
- Respect the user's culture and constraints; suggest, do not lecture.
- Times are local to the meeting; for remote attendees in other time zones, mention it but do not convert unless asked.
FILE:references/timeboxing-principles.md
# Timeboxing principles
## Every item needs an outcome
| Outcome | What the group does | Typical time | Notes |
|---|---|---|---|
| inform | Hear something, ask a clarifying question | 2 to 5 min | Better as a written pre-read. |
| update | Status from people or teams | 1 to 2 min per person | Usually async; keep only blockers live. |
| discuss | Explore a topic, share views, no decision yet | 10 to 20 min | Say what the discussion feeds (a decision next week, a doc). |
| brainstorm | Generate options | 15 to 30 min | Needs a framed question and a method (silent writing first). |
| review | Give feedback on a draft, design, or plan | 15 to 40 min | Send the material beforehand; agree what kind of feedback is wanted. |
| decide | Choose between options and commit | 10 to 30 min | Needs a written proposal, the options, the decision maker, and the decision rule. |
| break | Rest | 5 to 15 min | Every 60 to 90 minutes. |
If an item has no clear outcome, ask: "What will be different after this item?" If the answer is "nothing", cut it.
## How long things really take
- Introductions: about 1 minute per person for a new group, 0 for a regular team.
- Each person speaking once on a topic: about 1.5 to 2 minutes per person. Eight people, one round, is 15 minutes.
- Screen sharing and getting started: 2 to 3 minutes, more for remote meetings with new tools.
- A decision with real disagreement: double the estimate, or split it into "discuss" now and "decide" next time.
- Items late in the agenda run faster but worse: people rush or leave.
## Ordering
1. Start with a short framing: purpose and desired outcomes (2 to 5 min).
2. Decisions and the hardest topics in the first half, while energy and attendance are high.
3. Discussion and review in the middle.
4. Information and updates last or out of the meeting entirely.
5. End with next steps, owners, and dates (5 min), never with AOB.
Group items that need the same people, so others can leave early.
## Buffers and breaks
- Buffer: about 5 minutes per hour, 10 percent for workshops. The buffer is part of the slot, not extra time.
- Breaks: 5 to 10 minutes every 60 to 90 minutes; 15 for longer workshops. Remote meetings need breaks sooner.
- Back-to-back calendars: end 5 minutes early (25 or 50 minute meetings).
## Fit checks
- More than about 8 items per hour means each gets less than 7 minutes; merge or cut.
- Status and announcements above about a third of the time: move them to a pre-read.
- An item over 45 minutes: split it into steps with their own outcomes (for example "brainstorm 25" then "cluster and pick 20").
## The cost of a meeting
Person-hours = attendees x length in hours. A one-hour meeting with 9 people costs 9 working hours. Ask whether everyone needs to be there for every item.
FILE:references/facilitation-moves.md
# Facilitation moves that keep timeboxes
## Before the meeting
- Send the agenda and pre-reads 24 to 48 hours ahead, with the owner of each item.
- For every decision item, send a one-page proposal: the question, options, recommendation, and who decides.
- Ask people to add questions to the pre-read in writing, so the meeting starts from questions, not from a presentation.
## At the start
- State the purpose and the outcomes you need by the end.
- Name a timekeeper and agree on the signal (a message in the chat, a raised hand at 2 minutes left).
- Say how decisions are made today: the owner decides after input, majority vote, or consent ("any strong objection?").
## By item type
- Inform: "Two minutes, then one clarifying question each." Detailed questions go to the parking lot or to a follow-up chat.
- Update: Only blockers and asks live; everything else is in the pre-read.
- Discuss: Frame the question first ("What worries us about option B?"). Use a quick round so everyone speaks once before anyone speaks twice.
- Brainstorm: 3 to 5 minutes of silent writing first, then share, then cluster. Quantity before judgment.
- Review: Agree the kind of feedback: blocking issues, questions, or polish. Collect blocking issues first.
- Decide: Restate the proposal, list concerns, check for objections, then decide and say it out loud: "We decided X; owner Y; by date Z."
## When an item runs over
Make the choice visible instead of silently eating the next item:
1. "We have 2 minutes left on this. Do we extend by 5 minutes and cut something else, or park it?"
2. Park it: write it in the parking lot with an owner and a date.
3. Shrink the next item, never the decision items.
4. If the meeting goal is at risk, schedule a short follow-up with only the people needed.
## Remote and hybrid
- Everyone joins on their own device so remote people are equal.
- Use the chat for the speaking queue and for quick polls.
- Share the agenda on screen with the current item highlighted.
## At the end
- Read back decisions, owners, and dates (5 minutes).
- Review the parking lot: assign each item or drop it.
- Send notes within 24 hours.
FILE:templates/agenda.md
# Agenda input (for scripts/timebox_agenda.py)
```
Meeting: <name>
Start: <HH:MM>
Length: <minutes> # or End: <HH:MM>
Buffer: <minutes or percent, e.g. 5 or 10%>
Attendees: <number>
<minutes> | <topic> | <owner> | <inform | update | discuss | brainstorm | review | decide | break>
? | <topic with unknown duration> | <owner> | <outcome>
```
A Markdown table with the columns Min, Topic, Owner, Outcome works too, and so does JSON:
`{"meeting": "...", "start": "14:00", "length": 60, "buffer": 5, "attendees": 6, "items": [{"minutes": 10, "topic": "...", "owner": "...", "outcome": "decide"}]}`
---
# Agenda to send
**<Meeting name>**
<Day, date>, <start>-<end> (<length> min), <place or video link>
Attendees: <names or teams> | Facilitator: <name> | Timekeeper: <name> | Notes: <name>
**Purpose:** <one sentence: what must be true when we finish>
**Outcomes we need**
1. <decision or result>
2. <decision or result>
**Pre-read (by <day, time>)**
- <document> - <owner> - <what to look at, how long it takes to read>
**Schedule**
| Time | Min | Item | Owner | Outcome |
|---|---|---|---|---|
| <HH:MM-HH:MM> | <n> | <topic> | <owner> | <outcome> |
| <HH:MM-HH:MM> | <n> | Buffer and wrap-up: decisions, owners, dates | <facilitator> | |
**How we decide today:** <owner decides after input | consent | vote>
**Parking lot:** Off-topic points go to the parking lot and get an owner at the end.
**Moved out of this meeting**
- <item> -> <async update, other meeting, or a document>, owner <name>
FILE:examples/example-roadmap-meeting.md
# Example: fixing an overloaded one-hour planning meeting
## The request
"I have one hour with 9 people on Thursday at 14:00 to plan Q4. Here is my agenda. Will it fit? The budget split must be approved in this meeting, and the top 3 initiatives too."
Draft agenda (`agenda.txt`):
```
Meeting: Q4 roadmap planning
Start: 14:00
Length: 60
Buffer: 5
Attendees: 9
10 | Welcome and goals | Dana | inform
15 | Team status round (each team lead) | Leads | update
10 | Customer feedback highlights | Priya | inform
20 | Q4 priorities: pick top 3 initiatives | Lee | discuss
10 | Hiring plan for Q4 | | discuss
5 | Approve Q4 budget split | Dana | decide
? | AOB | |
```
## First run
```
$ python3 scripts/timebox_agenda.py agenda.txt
Q4 roadmap planning: 14:00-15:00 (60 min, buffer 5 min, 55 min for items)
Planned: 70 min
14:00-14:10 10 min Welcome and goals (Dana) [inform]
14:10-14:25 15 min Team status round (each team lead) (Leads) [update]
14:25-14:35 10 min Customer feedback highlights (Priya) [inform]
14:35-14:55 20 min Q4 priorities: pick top 3 initiatives (Lee) [discuss]
14:55-15:05 10 min Hiring plan for Q4 [discuss]
15:05-15:10 5 min Approve Q4 budget split (Dana) [decide]
15:10-15:10 ? min AOB [?]
14:55-15:00 5 min Buffer / wrap-up
Findings:
- [HIGH] Agenda needs 70 min but only 55 min are available after the 5-min buffer: overrun of 15 min (ends 15:10 instead of 14:55). Cut, shorten, or move items, or run with --fit.
- [WARN] 1 item(s) without a duration: AOB. Remaining time for them: 0 min.
- [WARN] "Hiring plan for Q4" has no owner; name who prepares and leads it.
- [WARN] "Approve Q4 budget split" is a decision with only 5 min; plan at least 10, or decide by a written proposal beforehand.
- [WARN] "Approve Q4 budget split" is a decision starting at 15:05, in the last quarter of the meeting; move it earlier while energy and attendance are high.
- [WARN] "AOB" has no owner; name who prepares and leads it.
- [WARN] "AOB" has no outcome; say whether it is to inform, discuss, decide, or brainstorm.
- [WARN] Status updates and announcements take 50% of the time; send them as a pre-read and keep the meeting for discussion and decisions.
- [INFO] Meeting cost: 9 people x 60 min = 9 person-hours.
```
Exit code 1: the agenda does not fit.
## What the numbers say
- 70 minutes of items (plus AOB) in a 55-minute window: 15 minutes over before AOB even starts.
- Half of the time is updates and announcements, while the two things that must be decided get 5 minutes (budget) and a "discuss" label (priorities), and the budget decision would start at 15:05, after the meeting should have ended.
- Running `--fit` alone would squeeze everything into 55 minutes but leave the budget decision last with 5 minutes. Fitting the numbers is not enough; the order and the purpose of items have to change.
## Changes
1. Team status round (15 min) and customer feedback highlights (10 min) become a written pre-read due Wednesday 12:00; questions go in comments.
2. The budget split moves to the start with 10 minutes and a proposal sent Monday, so the meeting only confirms or adjusts it.
3. "Q4 priorities" becomes a decision with 25 minutes: the outcome is a ranked top 3.
4. The hiring plan gets an owner (Priya) and a narrower question: which roles to open.
5. AOB is replaced by a 5-minute "next steps, owners, and parking lot" close.
6. Three people only needed for the status round can skip the meeting: 6 attendees instead of 9.
Revised agenda:
```
Meeting: Q4 roadmap planning
Start: 14:00
Length: 60
Buffer: 5
Attendees: 6
5 | Welcome and goals for today | Dana | inform
10 | Approve Q4 budget split (proposal sent Monday) | Dana | decide
25 | Q4 priorities: pick top 3 initiatives | Lee | decide
10 | Hiring plan for Q4: agree which roles to open | Priya | discuss
5 | Next steps, owners, and parking lot | Dana | inform
```
## Second run
```
$ python3 scripts/timebox_agenda.py agenda-revised.txt
Q4 roadmap planning: 14:00-15:00 (60 min, buffer 5 min, 55 min for items)
Planned: 55 min
14:00-14:05 5 min Welcome and goals for today (Dana) [inform]
14:05-14:15 10 min Approve Q4 budget split (proposal sent Monday) (Dana) [decide]
14:15-14:40 25 min Q4 priorities: pick top 3 initiatives (Lee) [decide]
14:40-14:50 10 min Hiring plan for Q4: agree which roles to open (Priya) [discuss]
14:50-14:55 5 min Next steps, owners, and parking lot (Dana) [inform]
14:55-15:00 5 min Buffer / wrap-up
Findings:
- [INFO] Meeting cost: 6 people x 60 min = 6 person-hours.
```
Exit code 0, no warnings, and the meeting costs 6 person-hours instead of 9.
## The agenda to send (filled template, shortened)
**Q4 roadmap planning**
Thursday, 14:00-15:00 (60 min), Room 3 and video link
Facilitator: Dana | Timekeeper: Lee | Notes: Priya
**Purpose:** Leave with an approved Q4 budget split and a ranked top 3 of initiatives.
**Pre-read (by Wednesday 12:00)**
- Team status one-pagers - each team lead - 5 min to read
- Customer feedback highlights - Priya - 5 min
- Budget split proposal - Dana - 10 min
**Schedule**
| Time | Min | Item | Owner | Outcome |
|---|---|---|---|---|
| 14:00-14:05 | 5 | Welcome and goals for today | Dana | inform |
| 14:05-14:15 | 10 | Approve Q4 budget split (proposal sent Monday) | Dana | decide |
| 14:15-14:40 | 25 | Q4 priorities: pick top 3 initiatives | Lee | decide |
| 14:40-14:50 | 10 | Hiring plan for Q4: agree which roles to open | Priya | discuss |
| 14:50-14:55 | 5 | Next steps, owners, and parking lot | Dana | inform |
| 14:55-15:00 | 5 | Buffer | | |
**How we decide today:** Dana decides the budget after input; the top 3 is chosen by consent (any strong objection?).
**Facilitation tips**
- Priorities: 3 minutes of silent dot voting on the candidate list, then discuss only the items ranked 2 to 5.
- If the budget item hits 10 minutes, park the open point with an owner and a 24-hour deadline instead of eating into priorities.
**Moved out of this meeting:** team status round and customer feedback (pre-read).
## Why this is a good answer
- It fixes the purpose and the order, not just the minutes.
- Both required decisions now happen in the first 40 minutes, with prepared proposals.
- The script confirms the fit, and the people who were only there for updates get their hour back.
FILE:scripts/timebox_agenda.py
#!/usr/bin/env python3
"""Timebox a meeting agenda: schedule every item, check the fit, and flag agenda smells.
Usage:
python3 timebox_agenda.py agenda.txt
python3 timebox_agenda.py agenda.txt --fit # scale items to fit the slot (5-minute steps)
python3 timebox_agenda.py agenda.json --json
cat agenda.txt | python3 timebox_agenda.py -
Text input: optional "Key: value" header lines, then one item per line as
minutes | topic | owner | outcome
Markdown table rows (| 10 | Topic | Owner | decide |) also work; header and
separator rows are skipped. Use ? for an unknown duration.
Header keys: Meeting, Start (HH:MM), Length (minutes) or End (HH:MM),
Buffer (minutes or percent, e.g. 5 or 10%), Attendees (number).
Outcomes: inform, update, discuss, brainstorm, review, decide, break.
JSON input: {"meeting": "...", "start": "14:00", "length": 60, "buffer": 5,
"attendees": 8, "items": [{"minutes": 10, "topic": "...", "owner": "...", "outcome": "decide"}]}
--fit trims status items (and items with no outcome) first, then scales everything; spare time goes to
decide and discuss items.
Checks: overrun (HIGH), missing durations, owners, or outcomes, late or rushed
decisions, oversized items, long stretches without a break, too many status
updates, too many items per hour, plus the meeting's cost in person-hours.
Exit code: 0 no HIGH findings, 1 HIGH found, 2 usage or input error.
Standard library only.
"""
import argparse
import json
import math
import re
import sys
OUTCOMES = {"inform", "update", "discuss", "brainstorm", "review", "decide", "break"}
ALIASES = {"decision": "decide", "info": "inform", "fyi": "inform", "status": "update", "updates": "update",
"discussion": "discuss", "ideas": "brainstorm", "ideation": "brainstorm", "feedback": "review",
"approve": "decide", "approval": "decide", "pause": "break", "coffee": "break"}
class InputError(Exception):
pass
def hhmm(text):
m = re.fullmatch(r"\s*(\d{1,2})[:.](\d{2})\s*", str(text))
if not m or int(m.group(1)) > 23 or int(m.group(2)) > 59:
raise InputError(f"time must be HH:MM, got {text!r}")
return int(m.group(1)) * 60 + int(m.group(2))
def fmt(minutes):
minutes = int(round(minutes)) % (24 * 60)
return f"{minutes // 60:02d}:{minutes % 60:02d}"
def parse_minutes(text):
t = str(text).strip().lower()
if t in ("", "?", "tbd", "-"):
return None
m = re.fullmatch(r"(\d+(?:\.\d+)?)\s*(m|min|mins|minutes?)?", t)
if m:
return float(m.group(1))
m = re.fullmatch(r"(\d+(?:\.\d+)?)\s*h(?:ours?)?(?:\s*(\d+)\s*m(?:in)?)?", t)
if m:
return float(m.group(1)) * 60 + float(m.group(2) or 0)
raise InputError(f"cannot read duration {text!r}")
def norm_outcome(text):
t = (text or "").strip().lower()
return ALIASES.get(t, t)
def parse_text(text):
meta, items, errors = {}, [], []
for no, raw in enumerate(text.replace("\r\n", "\n").split("\n"), 1):
line = raw.strip()
if not line or line.startswith("#"):
continue
hm = re.fullmatch(r"([A-Za-z ]{3,20}):\s*(.+)", line)
if hm and "|" not in line:
meta[hm.group(1).strip().lower()] = hm.group(2).strip()
continue
if "|" not in line:
errors.append(f"line {no}: expected 'minutes | topic | owner | outcome', got {line[:60]!r}")
continue
cells = [c.strip() for c in line.strip("|").split("|")]
if all(re.fullmatch(r":?-{2,}:?", c) for c in cells if c):
continue
if cells and cells[0].lower() in ("min", "mins", "minutes", "time", "duration"):
continue
cells += [""] * (4 - len(cells))
try:
minutes = parse_minutes(cells[0])
except InputError as e:
errors.append(f"line {no}: {e}")
continue
if not cells[1]:
errors.append(f"line {no}: item has no topic")
continue
items.append({"line": no, "minutes": minutes, "topic": cells[1], "owner": cells[2],
"outcome": norm_outcome(cells[3])})
return meta, items, errors
def parse_json(text):
data = json.loads(text)
if not isinstance(data, dict) or not isinstance(data.get("items"), list):
raise InputError("JSON input needs an object with an 'items' list")
meta = {k.lower(): str(v) for k, v in data.items() if k != "items" and v is not None}
items, errors = [], []
for i, it in enumerate(data["items"], 1):
try:
minutes = parse_minutes(it.get("minutes", ""))
except InputError as e:
errors.append(f"item {i}: {e}")
continue
if not str(it.get("topic", "")).strip():
errors.append(f"item {i}: item has no topic")
continue
items.append({"line": i, "minutes": minutes, "topic": str(it["topic"]).strip(),
"owner": str(it.get("owner", "") or "").strip(), "outcome": norm_outcome(it.get("outcome", ""))})
return meta, items, errors
def settings(meta):
start = hhmm(meta.get("start", "09:00"))
if "length" in meta:
length = parse_minutes(meta["length"])
if not length or length <= 0:
raise InputError("Length must be a positive number of minutes")
elif "end" in meta:
length = (hhmm(meta["end"]) - start) % (24 * 60)
if length == 0:
raise InputError("End must be after Start")
else:
raise InputError("header needs 'Length: <minutes>' or 'End: HH:MM'")
b = meta.get("buffer", "0").strip()
if b.endswith("%"):
buffer = 5 * round(length * float(b[:-1]) / 100 / 5) # percent buffers round to 5 minutes
else:
buffer = parse_minutes(b) or 0
if buffer >= length:
raise InputError("Buffer must be shorter than the meeting")
attendees = int(meta["attendees"]) if str(meta.get("attendees", "")).strip().isdigit() else None
return {"meeting": meta.get("meeting", meta.get("title", "Meeting")), "start": start, "length": length,
"buffer": buffer, "available": length - buffer, "attendees": attendees}
def fit(items, available):
"""Fit items into `available` minutes in 5-minute steps (missing durations count as 10, breaks kept).
When the agenda is too long, status items (inform, update, or no outcome) are trimmed first, down to half
their planned time; then every item is scaled proportionally. When it is too short, the
spare time goes to decide and discuss items first.
"""
fixed = sum(it["minutes"] or 0 for it in items if it["outcome"] == "break")
flex = [it for it in items if it["outcome"] != "break"]
room = available - fixed
if not flex or room < 5 * len(flex):
raise InputError(f"cannot fit {len(flex)} items into {room:g} minutes with at least 5 minutes each")
want = [max(5, 5 * round((it["minutes"] or 10) / 5)) for it in flex]
cur = list(want)
status = [i for i, it in enumerate(flex) if it["outcome"] in ("inform", "update", "")]
while sum(cur) > room:
cands = [i for i in status if cur[i] - 5 >= max(5, want[i] / 2)]
if not cands:
break
cur[max(cands, key=lambda k: cur[k])] -= 5
if sum(cur) > room:
scale = room / sum(cur)
ref = [c * scale for c in cur]
cur = [max(5, 5 * round(r / 5)) for r in ref]
while sum(cur) > room:
cands = [k for k in range(len(cur)) if cur[k] > 5]
cur[max(cands, key=lambda k: (cur[k] - ref[k], cur[k]))] -= 5
while sum(cur) + 5 <= room:
cur[min(range(len(cur)), key=lambda k: cur[k] - ref[k])] += 5
else:
prio = [i for i, it in enumerate(flex) if it["outcome"] in ("decide", "discuss")] or list(range(len(flex)))
while sum(cur) + 5 <= room:
cur[min(prio, key=lambda k: cur[k] / want[k])] += 5
out, it_new = [], iter(cur)
for it in items:
copy = dict(it)
copy["original"] = it["minutes"]
if it["outcome"] != "break":
copy["minutes"] = float(next(it_new))
out.append(copy)
return out
def analyze(items, cfg):
findings = []
def add(sev, msg):
findings.append({"severity": sev, "message": msg})
t = cfg["start"]
for it in items:
it["start"] = t
t += it["minutes"] or 0
it["end"] = t
planned = sum(it["minutes"] or 0 for it in items)
avail, end_slot = cfg["available"], cfg["start"] + cfg["length"]
unknown = [it for it in items if it["minutes"] is None]
if planned > avail:
add("HIGH", f"Agenda needs {planned:g} min but only {avail:g} min are available"
f"{' after the ' + format(cfg['buffer'], 'g') + '-min buffer' if cfg['buffer'] else ''}: overrun of {planned - avail:g} min"
f" (ends {fmt(cfg['start'] + planned)} instead of {fmt(end_slot - cfg['buffer'])}). Cut, shorten, or move items, or run with --fit.")
elif avail - planned > 0.15 * avail and not unknown:
add("INFO", f"{avail - planned:g} min unplanned; end early or give the main decision more time.")
if unknown:
add("WARN", f"{len(unknown)} item(s) without a duration: " + "; ".join(it["topic"][:40] for it in unknown)
+ f". Remaining time for them: {max(avail - planned, 0):g} min.")
for it in items:
name = f"\"{it['topic'][:50]}\""
if it["outcome"] != "break" and not it["owner"]:
add("WARN", f"{name} has no owner; name who prepares and leads it.")
if not it["outcome"]:
add("WARN", f"{name} has no outcome; say whether it is to inform, discuss, decide, or brainstorm.")
elif it["outcome"] not in OUTCOMES:
add("WARN", f"{name} has unknown outcome '{it['outcome']}' (use: {', '.join(sorted(OUTCOMES))}).")
m = it["minutes"] or 0
if it["outcome"] == "decide":
if it["minutes"] is not None and m < 10:
add("WARN", f"{name} is a decision with only {m:g} min; plan at least 10, or decide by a written proposal beforehand.")
if it["start"] - cfg["start"] >= 0.75 * cfg["length"]:
add("WARN", f"{name} is a decision starting at {fmt(it['start'])}, in the last quarter of the meeting; move it earlier while energy and attendance are high.")
if it["outcome"] in ("discuss", "brainstorm", "review") and it["minutes"] is not None and m < 5:
add("WARN", f"{name} is a {it['outcome']} item with {m:g} min; that is a mention, not a {it['outcome']}.")
if m > 45 and it["outcome"] != "break":
add("WARN", f"{name} is {m:g} min long; split it into parts with their own outcomes.")
stretch, stretch_start = 0, cfg["start"]
for it in items:
if it["outcome"] == "break":
stretch, stretch_start = 0, it["end"]
continue
stretch += it["minutes"] or 0
if stretch > 90:
add("WARN", f"{stretch:g} min without a break from {fmt(stretch_start)}; add a 5 to 10 min break about every 60 to 90 min.")
stretch = -10**9
status = sum(it["minutes"] or 0 for it in items if it["outcome"] in ("inform", "update"))
if planned and status / planned > 0.35:
add("WARN", f"Status updates and announcements take {round(100 * status / planned)}% of the time; send them as a pre-read and keep the meeting for discussion and decisions.")
work_items = [it for it in items if it["outcome"] != "break"]
hours = cfg["length"] / 60
if hours and len(work_items) / hours > 8:
add("WARN", f"{len(work_items)} items in {cfg['length']:g} min is more than 8 per hour; merge or drop items.")
if not any(it["outcome"] == "decide" for it in items):
add("INFO", "No decision items; check that this meeting needs to be live.")
cost = None
if cfg["attendees"]:
cost = round(cfg["attendees"] * cfg["length"] / 60, 1)
add("INFO", f"Meeting cost: {cfg['attendees']} people x {cfg['length']:g} min = {cost:g} person-hours.")
order = {"HIGH": 0, "WARN": 1, "INFO": 2}
findings.sort(key=lambda f: order[f["severity"]])
return {"meeting": cfg["meeting"], "start": fmt(cfg["start"]), "end": fmt(end_slot), "length": cfg["length"],
"buffer": cfg["buffer"], "available": avail, "planned": planned, "person_hours": cost,
"items": [{k: (fmt(v) if k in ("start", "end") else v) for k, v in it.items()} for it in items],
"findings": findings}
def main():
ap = argparse.ArgumentParser(description="Timebox a meeting agenda")
ap.add_argument("path", help="agenda text/Markdown/JSON file, or - for stdin")
ap.add_argument("--fit", action="store_true", help="scale item durations to fit the available time")
ap.add_argument("--json", action="store_true")
args = ap.parse_args()
try:
text = sys.stdin.read() if args.path == "-" else open(args.path, encoding="utf-8").read()
if not text.strip():
raise InputError("input is empty")
meta, items, errors = parse_json(text) if text.lstrip().startswith("{") else parse_text(text)
if errors:
raise InputError("could not read the agenda:\n " + "\n ".join(errors))
if not items:
raise InputError("no agenda items found (expected lines like '10 | Topic | Owner | decide')")
cfg = settings(meta)
before = sum(it["minutes"] or 0 for it in items)
n_unknown = sum(1 for it in items if it["minutes"] is None)
if args.fit:
items = fit(items, cfg["available"])
except FileNotFoundError:
print(f"error: file not found: {args.path}", file=sys.stderr)
return 2
except (InputError, json.JSONDecodeError, ValueError) as e:
print(f"error: {e}", file=sys.stderr)
return 2
rep = analyze(items, cfg)
rep["fitted_from"] = before if args.fit else None
if args.json:
print(json.dumps(rep, indent=2))
else:
print(f"{rep['meeting']}: {rep['start']}-{rep['end']} ({rep['length']:g} min, buffer {rep['buffer']:g} min, {rep['available']:g} min for items)")
if args.fit:
extra = f" plus {n_unknown} without a duration (counted as 10 min each)" if n_unknown else ""
print(f"Fitted: {before:g} min of items{extra} fitted into {rep['planned']:g} min (5-minute steps, breaks kept).")
print(f"Planned: {rep['planned']:g} min\n")
for it in rep["items"]:
mins = "?" if it["minutes"] is None else f"{it['minutes']:g}"
was = f" (was {it['original']:g})" if args.fit and it.get("original") not in (None, it["minutes"]) else ""
who = f" ({it['owner']})" if it["owner"] else ""
print(f" {it['start']}-{it['end']} {mins:>3} min{was:<10} {it['topic']}{who} [{it['outcome'] or '?'}]")
if rep["buffer"]:
print(f" {fmt(hhmm(rep['end']) - rep['buffer'])}-{rep['end']} {rep['buffer']:>3g} min Buffer / wrap-up")
print()
if rep["findings"]:
print("Findings:")
for f in rep["findings"]:
print(f"- [{f['severity']}] {f['message']}")
else:
print("Findings: none")
return 1 if any(f["severity"] == "HIGH" for f in rep["findings"]) else 0
if __name__ == "__main__":
sys.exit(main())Checks SRT and WebVTT subtitle files for broken or overlapping timings, numbering gaps, empty cues, reading speed above a characters-per-second limit, long or extra lines, cues that flash by or linger, unbalanced tags, and sound labels, then rewrites and retimes the problem cues and can shift the whole file. Use when a user shares subtitles or captions and asks "check my subtitles", "why are these captions hard to read?", "fix the timing", or "make these subtitles follow the guidelines".
---
name: srt-subtitle-quality-checker
description: Checks SRT and WebVTT subtitle files for broken or overlapping timings, numbering gaps, empty cues, reading speed above a characters-per-second limit, long or extra lines, cues that flash by or linger, unbalanced tags, and sound labels, then rewrites and retimes the problem cues and can shift the whole file. Use when a user shares subtitles or captions and asks "check my subtitles", "why are these captions hard to read?", "fix the timing", or "make these subtitles follow the guidelines".
---
# SRT Subtitle Quality Checker
You help video makers, translators, and accessibility teams ship subtitles that people can actually read. You find timing errors that break players, cues that appear too briefly or carry too much text, and lines that are too long, and you fix them by rewriting the text more concisely and adjusting times, without changing the meaning.
## Files in this skill
- `scripts/check_subtitles.py` - parses SRT and WebVTT, checks timing, numbering, reading speed, line length, line count, duration, gaps, tags, and sound labels, and can shift every timestamp (Python 3 standard library only)
- `references/subtitle-guidelines.md` - common limits for reading speed, line length, duration, gaps, line breaks, and when to use SDH labels
- `references/condensing-techniques.md` - how to shorten subtitle text without losing meaning, and how to retime cues
- `templates/subtitle-review.md` - the review report to return
- `examples/example-workshop-video.md` - a worked review of a short tutorial video
## Workflow
### 1. Collect the facts
Ask, or assume and say so:
- The file (SRT or VTT) and the video length; frame rate if known (24, 25, or 30 fps).
- Audience and style guide: general adult viewers, children, a platform guide, or the client's own limits.
- Subtitle type: same-language captions, translation, or SDH (for deaf and hard-of-hearing viewers, with sound labels).
- Whether you may rewrite text or only retime it.
### 2. Run the checker
```bash
python3 scripts/check_subtitles.py movie.srt
python3 scripts/check_subtitles.py movie.srt --max-cps 15 --max-line 37 # children or stricter guides
python3 scripts/check_subtitles.py movie.vtt --json
python3 scripts/check_subtitles.py movie.srt --shift -1200 > movie-synced.srt # whole file 1.2 s earlier
```
Defaults: 17 characters per second, 42 characters per line, 2 lines, 0.833 to 7 seconds on screen, 83 ms minimum gap. Pick limits from `references/subtitle-guidelines.md` to match the audience. If you cannot run the script, check the cues by hand: duration, characters divided by seconds, longest line, and overlaps with the next cue.
### 3. Fix in this order
1. HIGH findings first: overlaps, out-of-order cues, end before start, bad timestamps, empty cues. These break players or hide text.
2. Reading speed: extend the cue into free time before or after it (respect the minimum gap and shot changes the user mentions), otherwise condense the text with `references/condensing-techniques.md`, otherwise split the cue in two at a natural pause.
3. Line length and line count: rebreak at natural phrase boundaries; keep the top line shorter when possible; never break between an article and its noun.
4. Too short or too long on screen: merge very short cues with a neighbor, split long ones.
5. INFO items: renumber, balance tags, remove sound labels from non-SDH files, chain cues (0 ms gap) or leave a visible gap.
Rerun the checker on the corrected file until there are no HIGH findings and every WARN is fixed or explained.
### 4. Deliver
Fill in `templates/subtitle-review.md`: summary with counts before and after, the limits used, a table of changed cues (old and new times and text), and the full corrected file in the same format as the input. Follow the style of `examples/example-workshop-video.md`.
## Rules
- Never change the meaning, tone, or speaker of a line; condense filler, not content.
- Keep names, numbers, and technical terms exactly; if a term is wrong, flag it instead of guessing.
- Timing changes must stay in sync with the speech: extend a cue only into silence, by at most about 0.5 s before speech starts or 1 s after it ends, unless the user allows more.
- Keep the input format (SRT stays SRT, VTT stays VTT) and the original encoding (UTF-8).
- Say clearly which findings you could not fix without hearing the audio or seeing the video.
FILE:references/subtitle-guidelines.md
# Subtitle guidelines: common limits
Different broadcasters and platforms publish their own style guides. The values below are common starting points; always prefer the client's or platform's guide when one exists, and say which limits you used.
## Reading speed (characters per second, CPS)
CPS = visible characters in the cue (letters, digits, spaces, punctuation; not tags) divided by seconds on screen.
| Audience | Typical CPS limit |
|---|---|
| General adult viewers, same-language or translated | 15 to 17 |
| Fast-paced adult content where the guide allows it | up to 20 |
| Children (around 6 to 11 years) | 12 to 13 |
| Learners of the language, SDH with heavy sound labels | 13 to 15 |
Words per minute is an older measure; around 160 to 180 wpm corresponds to about 15 to 17 CPS in English.
## Line length and line count
- 37 to 42 characters per line for most Latin-script languages; 42 is the most common modern limit.
- Maximum 2 lines per cue. A third line covers the picture and is hard to read.
- Use one line when the text fits on one line; break into two only when needed.
## Duration on screen
- Minimum: about 5/6 of a second (20 frames at 24 fps), even for a single word, so the eye registers it.
- Maximum: about 7 seconds. Longer cues get reread and feel stuck; split them.
## Gaps between cues
- Chained cues: 0 ms gap is fine when the speech is continuous.
- Otherwise leave at least 2 frames (83 ms at 24 fps, 80 ms at 25 fps) so viewers notice that the text changed.
- Gaps between 1 ms and 2 frames cause a visible flicker and should be closed or widened.
## Line breaks
Break lines at natural linguistic units:
- After punctuation (comma, full stop) when possible.
- Before conjunctions (and, but, because) and prepositions.
- Never between an article and its noun, an adjective and its noun, a first and last name, or a verb and its auxiliary.
- Prefer a pyramid shape (shorter top line) when both breaks are equally good.
## Timing to speech and shots
- Start the cue when the speech starts (a few frames early is fine), end it no more than about 1 second after the speech ends.
- Avoid carrying a cue across a hard shot change; end it a couple of frames before the cut or start it after.
## SDH and captions
- Sound labels like [door slams] or (MUSIC) and speaker IDs belong in SDH/closed captions, not in standard translated subtitles.
- Keep labels short, lower case in square brackets is the most common modern style, and place them where the sound happens.
## Formatting
- Italics for off-screen voices, voice-over, songs, and foreign words, if the guide uses them. Every opening tag needs a closing tag in the same cue.
- No full stops at the end of single-line fragments is a style choice; follow the guide.
- SRT: cue number, timestamp line `00:01:02,500 --> 00:01:04,000`, text, blank line. VTT: `WEBVTT` header, optional cue id, timestamps with a dot `00:01:02.500`.
FILE:references/condensing-techniques.md
# Condensing and retiming subtitles
When a cue is too fast to read, try these in order. Stop as soon as the cue fits the limits.
## 1. Use free time first (no text change)
- Look at the gap before and after the cue. Extend the start up to about 0.5 s earlier if nobody else is speaking, and the end up to about 1 s later if the next cue starts later.
- Keep the minimum gap (2 frames) to the neighbors, or chain at 0 ms.
- Required time for a cue = characters / CPS limit. Example: 68 characters at 17 CPS need 4.0 s.
## 2. Cut filler, keep content
Remove words that the viewer hears or sees anyway:
- Hesitations and fillers: "well", "you know", "I mean", "um", "so, basically".
- Repetitions: "very, very" -> "very"; false starts.
- Greetings and names already obvious from the picture.
- Tag questions when tone is clear: "It's ready, isn't it?" -> "It's ready?" only if the meaning stays.
## 3. Use shorter forms
| Long | Short |
|---|---|
| we are going to | we'll |
| at this point in time | now |
| in order to | to |
| a large number of | many |
| it is not possible to | you can't |
| I would like to show you | let me show you |
| due to the fact that | because |
Contractions are usually fine in subtitles; follow the guide for formal content.
## 4. Simplify structure
- Turn passive into active: "The bowl was cleaned by the artist" -> "The artist cleans the bowl".
- Replace a clause with a single word when possible.
- Split one long sentence into two shorter cues at a natural pause.
## 5. Split or merge cues
- Split a cue longer than about 7 s, or one with a 3rd line, at a pause or punctuation; give each half time in proportion to its characters.
- Merge a cue under about 0.8 s with its neighbor when it is the same speaker and the result fits the limits.
## What never to cut
- Names, numbers, quantities, dates, technical terms, and negations ("not", "never").
- Words that carry the joke, the twist, or the emotion of the line.
- Information the viewer needs later in the video.
## Retiming the whole file
If every cue is early or late by the same amount, shift the whole file instead of editing cues:
```bash
python3 scripts/check_subtitles.py movie.srt --shift 1500 > movie-shifted.srt # 1.5 s later
python3 scripts/check_subtitles.py movie.srt --shift -800 > movie-shifted.srt # 0.8 s earlier
```
If the drift grows over time (in sync at the start, late at the end), the frame rate is probably wrong (for example 23.976 vs 25 fps); a constant shift will not fix it. Say so and ask for the frame rate of the video.
FILE:templates/subtitle-review.md
# Subtitle review: file_name
**Video:** video_title (duration, fps fps)
**Type:** translation / same-language captions / SDH
**Limits used:** cps chars/s, line chars/line, lines lines, min_dur to max_dur s on screen, gap ms minimum gap (source_of_limits)
## Summary
| | Before | After |
|---|---|---|
| Cues | | |
| HIGH | | |
| WARN | | |
| INFO | | |
| Average chars/s | | |
Main problems in one or two sentences.
## Changes
| Cue | Problem | Old time | New time | Old text | New text |
|---|---|---|---|---|---|
| | | | | | |
## Not fixed (needs audio or video)
- Cue ..., reason ...
## Notes for the editor
- Style choices made (italics, contractions, labels).
- Anything to check against the picture (shot changes, on-screen text).
## Corrected file
```
(full corrected SRT or VTT here, same format as the input)
```
FILE:examples/example-workshop-video.md
# Example: subtitle review of a short tutorial video
## User request
"Here are the English subtitles for our 25-second kintsugi workshop clip. Viewers say they can't keep up. Can you check them and fix what's wrong? General adult audience, normal subtitles, not SDH."
## Input (excerpt of interview.srt)
```
2
00:00:03,250 --> 00:00:04,000
Today we are going to look at how to repair a cracked ceramic bowl with gold.
3
00:00:03,900 --> 00:00:06,500
[HAMMER TAPPING]
First, clean the edges.
5
00:00:07,000 --> 00:00:16,000
<i>This technique is called kintsugi.
```
## Checker run
```
$ python3 scripts/check_subtitles.py interview.srt
interview.srt (SRT): 8 cues, 00:00:01,000 to 00:00:24,000, avg 16.3 chars/s, avg 2.37 s on screen
HIGH 3 WARN 9 INFO 3
HIGH overlap cue 3 @ 00:00:03,900: overlaps the previous cue by 100 ms; end the previous cue at 00:00:03,817
HIGH end-before-start cue 8 @ 00:00:21,000: ends at 00:00:20,500, not after its start
HIGH empty-cue cue 9 @ 00:00:22,000: cue has no visible text
WARN too-short cue 2 @ 00:00:03,250: on screen 0.75 s (minimum 0.83 s)
WARN reading-speed cue 2 @ 00:00:03,250: 102.7 chars/s (limit 17); needs 4.53 s on screen or about 64 fewer chars
...
```
## Review (filled template, shortened)
**Limits used:** 17 chars/s, 42 chars/line, 2 lines, 0.833 to 7 s, 83 ms gap (common general-audience values; no client guide given).
| | Before | After |
|---|---|---|
| Cues | 8 | 8 |
| HIGH | 3 | 0 |
| WARN | 9 | 0 |
| INFO | 3 | 0 |
Main problems: cue 2 carried a whole sentence for under a second, two cues had broken times, a 9-second cue lingered, and cue 7 had three long lines. All fixed by condensing filler, using the silence after the speech, and splitting one cue.
| Cue | Problem | Old time | New time | Old text | New text |
|---|---|---|---|---|---|
| 2 | 102.7 chars/s, 77-char line, overlap with 3 | 03,250-04,000 | 03,300-06,600 | Today we are going to look at how to repair a cracked ceramic bowl with gold. | Today we'll repair a cracked bowl / with lacquer and gold. |
| 3 | sound label in non-SDH file, overlap | 03,900-06,500 | 06,700-08,200 | [HAMMER TAPPING] / First, clean the edges. | First, clean the edges. |
| 4 (was 5) | 9 s on screen, unclosed italics, numbering gap | 07,000-16,000 | 08,300-11,000 | `<i>`This technique is called kintsugi. | `<i>`This technique is called kintsugi.`</i>` |
| 5 (was 6) | 0.4 s on screen | 16,000-16,400 | 16,000-16,900 | Okay. | Okay. |
| 6 and 7 (was 7) | 56 chars/s, 3 lines, 67-char line | 17,000-19,000 | 17,000-21,000 and 21,100-22,800 | You will need lacquer, a fine brush, and gold powder, and patience, / lots of patience, / because each layer must dry. | You'll need lacquer, a fine brush, / gold powder, and lots of patience. + Each layer must dry first. |
| 8 (was 8) | ends before it starts | 21,000-20,500 | 22,900-24,000 | Let's start. | Let's start. |
| 9 | empty cue | 22,000-24,000 | removed | | |
Note: the old cue 2 was shortened ("look at how to", "ceramic") only where the picture already shows it; "lacquer" was added because the speaker names it a few seconds later and the viewer needs it.
**Not fixed:** none, but please confirm against the video that the speaker pauses after "gold" (cue 2 now runs 2.6 s longer, into what looks like silence) and that "Let's start." is spoken at about 22.9 s.
Corrected file rechecked:
```
$ python3 scripts/check_subtitles.py interview-fixed.srt
interview-fixed.srt (SRT): 8 cues, 00:00:01,000 to 00:00:24,000, avg 14.5 chars/s, avg 2.17 s on screen
HIGH 0 WARN 0 INFO 0
no problems found
```
FILE:scripts/check_subtitles.py
#!/usr/bin/env python3
"""Check SRT or WebVTT subtitle files for timing and readability problems.
Usage:
python3 check_subtitles.py FILE [FILE ...] [options]
python3 check_subtitles.py - < movie.srt (read stdin)
Options:
--max-cps N maximum characters per second (default 17)
--max-line N maximum characters per line (default 42)
--max-lines N maximum lines per cue (default 2)
--min-dur S minimum cue duration in seconds (default 0.833, 20 frames at 24 fps)
--max-dur S maximum cue duration in seconds (default 7.0)
--min-gap MS minimum gap between cues in milliseconds (default 83, 2 frames at 24 fps)
--shift MS print the file with every timestamp shifted by MS (may be negative) and exit
--json print findings as JSON
--fail-on LEVEL exit 1 if any finding is at least LEVEL: high (default), warn, info
Findings: HIGH = broken or overlapping timing, unreadable numbering, empty cue;
WARN = reading speed, line length, too many lines, too short or long on screen;
INFO = tiny gaps, unbalanced tags, hearing-impaired brackets, statistics.
Exit codes: 0 ok, 1 findings at or above --fail-on, 2 usage or parse error.
Python 3 standard library only.
"""
import json
import math
import re
import sys
TS = re.compile(r"^\s*(\d{1,2}:)?(\d{1,2}):(\d{2})[,.](\d{1,3})\s*-->\s*(\d{1,2}:)?(\d{1,2}):(\d{2})[,.](\d{1,3})(.*)$")
TAG = re.compile(r"</?[^>]+>|\{\\[^}]*\}")
LEVELS = {"info": 0, "warn": 1, "high": 2}
def to_ms(h, m, s, ms):
h = int(h[:-1]) if h else 0
return ((h * 60 + int(m)) * 60 + int(s)) * 1000 + int(ms.ljust(3, "0"))
def fmt(ms, vtt=False):
sign = "-" if ms < 0 else ""
ms = abs(ms)
h, rem = divmod(ms, 3600000)
m, rem = divmod(rem, 60000)
s, rem = divmod(rem, 1000)
sep = "." if vtt else ","
return f"{sign}{h:02d}:{m:02d}:{s:02d}{sep}{rem:03d}"
def parse(text):
"""Return (cues, errors, is_vtt). Each cue: dict(index, start, end, lines, line_no)."""
text = text.replace("\r\n", "\n").replace("\r", "\n").lstrip("\ufeff")
is_vtt = text.lstrip().startswith("WEBVTT")
blocks = re.split(r"\n\s*\n", text.strip("\n"))
cues, errors = [], []
line_no = 1
for block in blocks:
lines = block.split("\n")
start_line = line_no
line_no += len(lines) + 1
if is_vtt and (lines[0].startswith("WEBVTT") or lines[0].startswith("NOTE") or lines[0].startswith("STYLE") or lines[0].startswith("REGION")):
continue
ts_i = next((i for i, ln in enumerate(lines[:3]) if "-->" in ln), None)
if ts_i is None:
errors.append({"level": "high", "rule": "unparsed-block", "line": start_line,
"message": f"block without a timestamp line: {lines[0][:50]!r}"})
continue
m = TS.match(lines[ts_i])
if not m:
errors.append({"level": "high", "rule": "bad-timestamp", "line": start_line + ts_i,
"message": f"cannot read timestamp {lines[ts_i].strip()!r} (expected 00:01:02,500 --> 00:01:04,000)"})
continue
g = m.groups()
index = lines[0].strip() if ts_i == 1 else None
cues.append({"index": index, "start": to_ms(*g[0:4]), "end": to_ms(*g[4:8]),
"lines": [ln for ln in lines[ts_i + 1:]], "line_no": start_line})
return cues, errors, is_vtt
def visible(line):
return TAG.sub("", line).strip()
def check(cues, errors, opts, is_vtt):
out = list(errors)
def add(level, rule, cue, msg):
out.append({"level": level, "rule": rule, "line": cue["line_no"],
"cue": cue["index"] or "-", "time": fmt(cue["start"], is_vtt), "message": msg})
if not is_vtt:
expected = 1
for c in cues:
if c["index"] is None or not c["index"].isdigit():
add("high", "bad-number", c, f"cue number missing or not a number: {c['index']!r}")
elif int(c["index"]) != expected:
add("warn", "numbering", c, f"cue number {c['index']} but expected {expected} (renumber the file)")
expected = int(c["index"])
expected += 1
prev = None
for c in cues:
dur = (c["end"] - c["start"]) / 1000
text_lines = [visible(ln) for ln in c["lines"] if visible(ln)]
chars = sum(len(ln) for ln in text_lines)
if c["end"] <= c["start"]:
add("high", "end-before-start", c, f"ends at {fmt(c['end'], is_vtt)}, not after its start")
if not text_lines:
add("high", "empty-cue", c, "cue has no visible text")
if prev is not None:
gap = c["start"] - prev["end"]
if c["start"] < prev["start"]:
add("high", "out-of-order", c, f"starts before the previous cue ({fmt(prev['start'], is_vtt)}); sort by time")
elif gap < 0:
add("high", "overlap", c, f"overlaps the previous cue by {-gap} ms; end the previous cue at {fmt(c['start'] - opts['min_gap'], is_vtt)}")
elif 0 < gap < opts["min_gap"]:
add("info", "tiny-gap", c, f"only {gap} ms after the previous cue; use 0 ms (chained) or at least {opts['min_gap']} ms so the change is visible")
if c["end"] > c["start"] and text_lines:
if dur < opts["min_dur"]:
add("warn", "too-short", c, f"on screen {dur:.2f} s (minimum {opts['min_dur']:.2f} s)")
if dur > opts["max_dur"]:
add("warn", "too-long", c, f"on screen {dur:.2f} s (maximum {opts['max_dur']:.1f} s); split it or shorten the time")
cps = chars / dur if dur > 0 else float("inf")
if cps > opts["max_cps"]:
need = chars / opts["max_cps"]
cut = max(1, math.ceil(chars - opts["max_cps"] * dur))
add("warn", "reading-speed", c, f"{cps:.1f} chars/s (limit {opts['max_cps']:g}); needs {need:.2f} s on screen or about {cut} fewer chars")
for ln in text_lines:
if len(ln) > opts["max_line"]:
add("warn", "line-length", c, f"line has {len(ln)} chars (limit {opts['max_line']}): {ln[:60]!r}")
if len(text_lines) > opts["max_lines"]:
add("warn", "too-many-lines", c, f"{len(text_lines)} lines (limit {opts['max_lines']})")
raw = " ".join(c["lines"])
for tag in ("i", "b", "u"):
if len(re.findall(rf"<{tag}>", raw, re.I)) != len(re.findall(rf"</{tag}>", raw, re.I)):
add("info", "unbalanced-tag", c, f"<{tag}> tags are not balanced in this cue")
if re.search(r"\[[^\]]+\]|\([A-Z][A-Z ]+\)", raw):
add("info", "sdh-label", c, "contains a sound or speaker label in brackets; keep it only in SDH/CC files")
prev = c
return out
def stats(cues, is_vtt=False):
if not cues:
return {}
total_chars = sum(len(visible(ln)) for c in cues for ln in c["lines"])
total_time = sum(max(c["end"] - c["start"], 0) for c in cues) / 1000
return {"cues": len(cues), "first": fmt(cues[0]["start"], is_vtt), "last_end": fmt(max(c["end"] for c in cues), is_vtt),
"avg_cps": round(total_chars / total_time, 1) if total_time else None,
"avg_duration_s": round(total_time / len(cues), 2)}
def shift_text(text, delta):
def rep(m):
g = m.groups()
vtt = "." in m.group(0).split("-->")[0]
a = max(to_ms(*g[0:4]) + delta, 0)
b = max(to_ms(*g[4:8]) + delta, 0)
return f"{fmt(a, vtt)} --> {fmt(b, vtt)}{g[8]}"
return "\n".join(TS.sub(rep, ln) if "-->" in ln else ln
for ln in text.replace("\r\n", "\n").split("\n"))
def main(argv):
opts = {"max_cps": 17.0, "max_line": 42, "max_lines": 2, "min_dur": 0.833, "max_dur": 7.0,
"min_gap": 83, "json": False, "fail_on": "high", "shift": None}
files = []
it = iter(argv)
try:
for a in it:
if a == "--json":
opts["json"] = True
elif a in ("--max-cps", "--min-dur", "--max-dur"):
opts[a[2:].replace("-", "_")] = float(next(it))
elif a in ("--max-line", "--max-lines", "--min-gap", "--shift"):
opts[a[2:].replace("-", "_")] = int(next(it))
elif a == "--fail-on":
opts["fail_on"] = next(it).lower()
if opts["fail_on"] not in LEVELS:
raise ValueError("--fail-on must be high, warn, or info")
elif a in ("-h", "--help"):
print(__doc__)
return 0
elif a.startswith("--"):
raise ValueError(f"unknown option {a}")
else:
files.append(a)
except (StopIteration, ValueError) as e:
print(f"error: {e or 'option needs a value'}", file=sys.stderr)
return 2
if not files:
print("error: give at least one .srt or .vtt file, or - for stdin", file=sys.stderr)
return 2
report, worst = [], -1
for f in files:
try:
text = sys.stdin.read() if f == "-" else open(f, encoding="utf-8-sig").read()
except (OSError, UnicodeDecodeError) as e:
print(f"error: cannot read {f}: {e}", file=sys.stderr)
return 2
if opts["shift"] is not None:
sys.stdout.write(shift_text(text, opts["shift"]))
return 0
cues, errors, is_vtt = parse(text)
if not cues:
print(f"error: {f}: no subtitle cues found (is this an SRT or WebVTT file?)", file=sys.stderr)
return 2
findings = check(cues, errors, opts, is_vtt)
for x in findings:
worst = max(worst, LEVELS[x["level"]])
report.append({"file": f, "format": "vtt" if is_vtt else "srt", "stats": stats(cues, is_vtt), "findings": findings})
if opts["json"]:
print(json.dumps(report, indent=2))
else:
for r in report:
counts = {lv: sum(1 for x in r["findings"] if x["level"] == lv) for lv in ("high", "warn", "info")}
st = r["stats"]
print(f"{r['file']} ({r['format'].upper()}): {st['cues']} cues, {st['first']} to {st['last_end']}, "
f"avg {st['avg_cps']} chars/s, avg {st['avg_duration_s']} s on screen")
print(f" HIGH {counts['high']} WARN {counts['warn']} INFO {counts['info']}")
for x in sorted(r["findings"], key=lambda x: (-LEVELS[x["level"]], x["line"])):
where = f"cue {x.get('cue', '-')} @ {x['time']}" if "time" in x else f"line {x['line']}"
print(f" {x['level'].upper():<4} {x['rule']:<16} {where}: {x['message']}")
if not r["findings"]:
print(" no problems found")
return 1 if worst >= LEVELS[opts["fail_on"]] else 0
if __name__ == "__main__":
sys.exit(main(sys.argv[1:]))Builds fixed-rate loan and mortgage amortization schedules, shows how much of each payment goes to interest, and compares payoff strategies such as extra monthly payments, one-off lump sums, a higher fixed payment, or a shorter term, with interest saved, months saved, payoff date, and an APR estimate when fees apply. Use when a user asks "how much interest will I pay?", "should I overpay my mortgage?", "what if I pay 200 more a month?", or wants a payoff plan for a car, student, or home loan.
--- name: loan-amortization-planner description: Builds fixed-rate loan and mortgage amortization schedules, shows how much of each payment goes to interest, and compares payoff strategies such as extra monthly payments, one-off lump sums, a higher fixed payment, or a shorter term, with interest saved, months saved, payoff date, and an APR estimate when fees apply. Use when a user asks "how much interest will I pay?", "should I overpay my mortgage?", "what if I pay 200 more a month?", or wants a payoff plan for a car, student, personal, or home loan. Includes a tested stdlib Python calculator. --- # Loan Amortization Planner You help people understand what a loan really costs and how to pay it off faster in a way that fits their budget. You run the numbers precisely, explain them in plain language, compare realistic options side by side, and point out the trade-offs (emergency savings, other debts, prepayment rules) before anyone sends extra money to a lender. ## Files in this skill - `scripts/amortize.py` - monthly amortization for fixed-rate loans with extra monthly payments, lump sums, a fixed payment, fees and an APR estimate, a yearly table, and a CSV schedule (Python 3 standard library only) - `references/amortization-basics.md` - how amortization works, the formulas, and the terms people confuse (rate vs APR, term vs payoff date) - `references/prepayment-decision-guide.md` - when overpaying makes sense, what to check in the loan contract, and the order of priorities - `templates/loan-plan.md` - the report to return - `examples/example-mortgage-overpayment.md` - a worked comparison of three overpayment strategies on a mortgage ## Workflow ### 1. Collect the facts Ask, or assume and say so: - Amount still owed (or amount to borrow), interest rate, remaining term, and the current monthly payment if known. - Fixed or variable rate, and when a fixed period ends. - Fees: arrangement or origination fees, and any prepayment penalty or yearly overpayment limit. - What the user can afford: an extra amount per month, an expected bonus or one-off sum, and their emergency savings and other debts. - The goal: lowest total interest, being debt-free by a date, or lowest monthly payment. ### 2. Run the baseline ```bash python3 scripts/amortize.py --principal 300000 --rate 4.5 --years 30 --start 2026-12 python3 scripts/amortize.py --principal 18500 --rate 7.9 --months 48 --fee 400 # car loan with a fee: APR estimate ``` Check that the computed payment matches the lender's statement within a few cents. If it does not, the loan probably has a different day count, fees in the balance, or insurance in the payment: ask, or use `--payment` with the real amount. ### 3. Run the scenarios ```bash python3 scripts/amortize.py --principal 300000 --rate 4.5 --years 30 --extra 200 python3 scripts/amortize.py --principal 300000 --rate 4.5 --years 30 --lump 12:10000 --lump 24:10000 python3 scripts/amortize.py --principal 300000 --rate 4.5 --years 30 --payment 2000 --yearly python3 scripts/amortize.py --principal 300000 --rate 4.5 --years 20 # shorter term python3 scripts/amortize.py ... --schedule plan.csv --json ``` Compare 2 to 4 options that the user can really afford. For each, record the monthly outlay, payoff date, total interest, interest saved, and months saved. If you cannot run the script, use the formulas in `references/amortization-basics.md` and say the numbers are approximate. ### 4. Check the decision, not just the math Go through `references/prepayment-decision-guide.md`: emergency fund first, higher-rate debts first, prepayment penalties and limits, tax effects to check locally, and the after-tax return of the alternative (saving or investing). Never present overpaying as always right. ### 5. Deliver Fill in `templates/loan-plan.md`: the loan in one line, the baseline, a comparison table of scenarios, a recommendation with reasons, what to ask the lender, and next steps. Follow the style of `examples/example-mortgage-overpayment.md`. ## Rules - Show every number with its currency and round money to cents; label estimates as estimates. - The calculator assumes a fixed rate, monthly payments, and extra payments applied to principal straight away with the payment unchanged (the term shortens). Say so, and flag variable rates or lenders that recalculate the payment instead. - Do not give personal investment or tax advice; explain the trade-off and tell the user what to check with the lender or a licensed adviser. - If a payment does not even cover the interest, say so clearly and stop: the balance would grow. - Keep the explanation short and concrete: "every 100 of extra saves about 120 of interest" beats a paragraph of theory. FILE:references/amortization-basics.md # Amortization basics ## How a fixed-rate loan is paid off Each month the lender charges interest on the balance still owed. Your payment first covers that interest; the rest reduces the balance (principal). Because the balance falls, the interest part shrinks every month and the principal part grows, while the payment stays the same. Early payments are mostly interest; late payments are mostly principal. ## Formulas - Monthly rate: r = annual rate / 12 (for example 4.5% -> 0.00375). - Payment for a loan P over n months: M = P * r / (1 - (1 + r)^-n). With r = 0: M = P / n. - Interest this month: balance * r. Principal this month: M - interest. - Total interest: sum of the interest parts, or approximately M * n - P. - Months to pay off with payment M: n = -ln(1 - P * r / M) / ln(1 + r). If P * r >= M the loan never ends. Quick checks: 200,000 at 6% for 30 years -> 1,199.10 per month. 300,000 at 4.5% for 30 years -> 1,520.06 per month. ## Why extra payments save so much An extra payment reduces the balance immediately, so every later month charges interest on a smaller amount. The saving is roughly the extra amount times the interest it would have carried for the rest of the loan. That is why money paid early in the loan saves more than the same money paid late. ## Terms people confuse | Term | Meaning | |---|---| | Nominal rate | The yearly rate in the contract, used for the monthly interest. | | APR | Annual percentage rate including fees; compares offers. A loan with a lower rate but high fees can have a higher APR. | | Term | The planned length (e.g. 30 years). | | Payoff date | When the balance actually reaches zero; earlier with extra payments. | | Reduce term vs reduce payment | After an overpayment, some lenders keep the payment and shorten the term (saves more interest), others recalculate a lower payment. Ask which. | | Principal | The amount still owed, not counting future interest. | | Amortization schedule | The month-by-month table of payment, interest, principal, and balance. | ## Rounding and small differences Lenders round to cents each month and may use daily interest (actual days / 365) instead of rate / 12. Results can differ by a few cents per month and a few units in total interest. The last payment usually absorbs the rounding difference. ## Limits of a fixed-rate calculator - Variable or tracker rates: run scenarios at the current rate and at +1 and +2 percentage points. - Interest-only periods, balloon payments, payment holidays: model them separately or by hand. - Insurance, escrow, or account fees inside the monthly payment are not interest; remove them before comparing. FILE:references/prepayment-decision-guide.md # Should I pay off my loan faster? A decision guide Overpaying a loan gives a guaranteed, risk-free "return" equal to the loan's interest rate. That is good, but it is not always the best use of spare money. Work through these steps in order. ## 1. Safety first - Emergency fund: keep about 3 to 6 months of essential costs in easy-access savings before overpaying. Money sent to a lender is hard to get back. - Stable income: if a job change, parental leave, or a big expense is coming, build cash first. ## 2. Pay the most expensive debt first Rank debts by interest rate (avalanche method): credit cards and overdrafts, then personal and car loans, then student loans and mortgages. Overpay the highest rate first. If motivation is the problem, paying the smallest balance first (snowball) can be fine, but say what it costs. ## 3. Read the contract - Prepayment penalty or early repayment charge (often a percent of the amount overpaid, mostly during a fixed-rate period). - Yearly overpayment allowance (for example 10 percent of the balance per year without a fee). - Does an overpayment reduce the term or the payment? Can you choose? - Minimum overpayment amounts and how to make one (online, by phone, written instruction). ## 4. Compare with the alternative - Employer retirement matching is usually better than any overpayment: take the full match first. - Compare the loan rate with the after-tax, after-fee return you could get on savings. A 2% mortgage versus a 4% savings account: saving wins on paper. A 7% car loan versus a 3% savings account: overpaying wins. - Investing may beat a low-rate loan over long periods but carries risk; the overpayment return is certain. - Tax: in some countries mortgage interest is tax-deductible, which lowers the effective rate. Check locally; do not assume. ## 5. Pick a strategy | Strategy | Good when | |---|---| | Fixed extra per month | Steady income; builds a habit; saves the most for the same total if started early. | | Lump sums (bonus, tax refund) | Irregular income; keeps flexibility during the year. | | Higher fixed payment | Wants a clear debt-free date. | | Refinance to a shorter term | A lower rate is available and the higher payment is affordable; compare fees. | | Keep cash, overpay later | Fixed period ending soon (overpay without penalty after), or emergency fund not full. | ## 6. Review Re-run the numbers once a year, when the rate changes, or when income changes. Remind the user that overpaying can always be paused; a higher contractual payment usually cannot. FILE:templates/loan-plan.md # Loan plan: loan_name **Loan:** balance at rate% (fixed_or_variable), remaining_term left, payment payment per month, first payment in this plan start_month **Fees and rules:** fees; prepayment limit or penalty: prepayment_rules **Goal:** goal **Budget for extra payments:** extra_budget ## Baseline - Monthly payment: - Total interest from now: - Payoff date: - First payment split: interest ... / principal ... ## Scenarios | Option | Monthly outlay | Payoff date | Total interest | Interest saved | Months saved | |---|---|---|---|---|---| | Baseline | | | | - | - | | A: | | | | | | | B: | | | | | | | C: | | | | | | ## Recommendation Option ... because: 1. 2. 3. Trade-offs to keep in mind: ## Before you overpay - [ ] Emergency fund covers ... months - [ ] No higher-rate debt left (or it comes first) - [ ] Prepayment penalty / allowance checked with the lender - [ ] Overpayment set to reduce the term (or payment, if preferred) ## Questions for the lender 1. 2. ## Next steps 1. 2. *Assumptions: fixed rate, monthly payments, extra payments applied to principal immediately. These are estimates, not financial advice.* FILE:examples/example-mortgage-overpayment.md # Example: comparing three mortgage overpayment strategies ## User request "We have 320,000 left on our mortgage at 4.1% fixed with 25 years to go, first payment of this plan in December 2026. We could pay about 300 extra a month, or we get a 15,000 bonus each year for the next three years. A friend says we should just switch to 2,400 a month. What's best? We have 6 months of expenses saved and no other debt. The bank allows 10% overpayment per year without fees." ## Commands ```bash python3 scripts/amortize.py --principal 320000 --rate 4.1 --years 25 --start 2026-12 python3 scripts/amortize.py --principal 320000 --rate 4.1 --years 25 --start 2026-12 --extra 300 python3 scripts/amortize.py --principal 320000 --rate 4.1 --years 25 --start 2026-12 --lump 12:15000 --lump 24:15000 --lump 36:15000 python3 scripts/amortize.py --principal 320000 --rate 4.1 --years 25 --start 2026-12 --payment 2400 ``` Output (shortened): ``` Baseline: 1,706.80 per month, interest 192,038.38, total paid 512,038.38, paid off 2051-11 (300 payments) Plan (+300.00/month): paid off 2046-02 (231 payments), interest 143,068.39 Saves 48,969.99 interest and 69 months (5 y 9 m); every 100 of extra saves 70.97 Plan (lump sums 15,000.00 in payment 12, 15,000.00 in payment 24, 15,000.00 in payment 36): paid off 2046-11 (240 payments), interest 133,018.73 Saves 59,019.65 interest and 60 months (5 y 0 m); every 100 of extra saves 131.15 Plan (fixed payment 2,400.00): paid off 2041-10 (179 payments), interest 107,805.37 Saves 84,233.01 interest and 121 months (10 y 1 m) ``` ## Loan plan (filled template, shortened) **Loan:** 320,000.00 at 4.1% fixed, 25 years left, payment 1,706.80 per month, first payment December 2026 **Fees and rules:** overpayments up to 10% of the balance per year are free **Goal:** pay less interest without losing flexibility | Option | Monthly outlay | Payoff date | Total interest | Interest saved | Months saved | |---|---|---|---|---|---| | Baseline | 1,706.80 | 2051-11 | 192,038.38 | - | - | | A: +300 per month | 2,006.80 | 2046-02 | 143,068.39 | 48,969.99 | 69 | | B: 15,000 bonus in years 1 to 3 | 1,706.80 + 45,000 total | 2046-11 | 133,018.73 | 59,019.65 | 60 | | C: fixed 2,400 per month | 2,400.00 | 2041-10 | 107,805.37 | 84,233.01 | 121 | **Recommendation:** Option B, plus A if the budget allows. The bonuses are paid early in the loan, so each 100 overpaid saves about 131 of interest, almost twice the rate of the monthly plan (about 71 per 100, because much of that money is paid in later years). It needs no change to the monthly budget, and each 15,000 is within the free 10% allowance (32,000 in year 1). Option C saves the most because it sends about 124,000 of extra money over 15 years, but it raises the fixed outlay by 693 per month; do it only if that fits comfortably, and set it up as a voluntary overpayment rather than a new contract payment so it can be paused. **Trade-offs:** money overpaid is hard to get back; keep the 6-month emergency fund untouched. If your savings account pays more than 4.1% after tax, saving could beat overpaying on paper; check the rate and any tax rules locally. **Questions for the lender:** 1. Will overpayments reduce the term with the payment unchanged (assumed above), or recalculate the payment? 2. Does the 10% allowance reset per calendar year or per loan year? *Assumptions: fixed rate for the whole term, monthly payments, overpayments applied to principal immediately. Estimates, not financial advice.* FILE:scripts/amortize.py #!/usr/bin/env python3 """Fixed-rate loan amortization with extra payments and scenario comparison. Usage: python3 amortize.py --principal 250000 --rate 4.2 --years 30 [options] python3 amortize.py --principal 18000 --rate 7.9 --months 60 --extra 150 Options: --principal N amount borrowed (required) --rate PCT nominal annual interest rate in percent, e.g. 4.2 (required; 0 allowed) --years N | --months N term (one of them is required) --start YYYY-MM month of the first payment (default: next month) --extra N extra amount added to every monthly payment --extra-from N payment number where --extra starts (default 1) --lump N:AMOUNT one-off extra payment in payment number N (repeatable) --payment N pay this fixed monthly amount instead of the computed one --fee N one-off upfront fees, used for the cost and APR estimate --schedule FILE write the full monthly schedule as CSV --yearly print a year-by-year table --json print the result as JSON Extra payments are assumed to reduce principal immediately with no prepayment penalty and the same monthly payment (term shortens). Rates are nominal annual, compounded monthly (rate / 12 per month). Money is rounded to cents each month. Exit codes: 0 ok, 2 usage or input error. Python 3 standard library only. """ import argparse import csv import datetime as dt import json import sys def monthly_payment(principal, annual_rate, n): r = annual_rate / 100 / 12 if r == 0: return principal / n return principal * r / (1 - (1 + r) ** -n) def add_months(ym, k): y, m = ym m0 = m - 1 + k return (y + m0 // 12, m0 % 12 + 1) def schedule(principal, annual_rate, n, payment=None, extra=0.0, extra_from=1, lumps=None): r = annual_rate / 100 / 12 base = round(payment if payment else monthly_payment(principal, annual_rate, n), 2) lumps = lumps or {} bal = round(principal, 2) rows = [] k = 0 while bal > 0.005: k += 1 if k > 1200: raise ValueError("the loan is not paid off within 100 years; the payment is too small for the interest") interest = round(bal * r, 2) if base + (extra if k >= extra_from else 0) <= interest and not lumps.get(k): raise ValueError(f"payment {base:.2f} does not cover the monthly interest {interest:.2f}; the balance would grow") sched = min(base, bal + interest) if payment is None and k == n: sched = bal + interest # last scheduled payment absorbs the rounding difference ext = min(extra if k >= extra_from else 0.0, max(bal + interest - sched, 0)) lump = min(lumps.get(k, 0.0), max(bal + interest - sched - ext, 0)) principal_paid = round(sched + ext + lump - interest, 2) bal = round(bal - principal_paid, 2) rows.append({"n": k, "payment": round(sched, 2), "extra": round(ext + lump, 2), "interest": interest, "principal": principal_paid, "balance": max(bal, 0.0)}) return base, rows def summarize(rows): return {"months": len(rows), "total_interest": round(sum(x["interest"] for x in rows), 2), "total_paid": round(sum(x["payment"] + x["extra"] for x in rows), 2), "total_extra": round(sum(x["extra"] for x in rows), 2)} def apr_estimate(principal, fee, payment, n): """Annual rate that makes the payments worth principal - fee (bisection).""" target = principal - fee lo, hi = 0.0, 1.0 for _ in range(100): mid = (lo + hi) / 2 r = mid / 12 pv = payment * n if r == 0 else payment * (1 - (1 + r) ** -n) / r lo, hi = (mid, hi) if pv > target else (lo, mid) return round(lo * 100, 3) def fmt_ym(ym): return f"{ym[0]:04d}-{ym[1]:02d}" def main(argv): p = argparse.ArgumentParser(add_help=True, description="Loan amortization with extra payments") p.add_argument("--principal", type=float, required=True) p.add_argument("--rate", type=float, required=True) g = p.add_mutually_exclusive_group(required=True) g.add_argument("--years", type=float) g.add_argument("--months", type=int) p.add_argument("--start") p.add_argument("--extra", type=float, default=0.0) p.add_argument("--extra-from", type=int, default=1) p.add_argument("--lump", action="append", default=[]) p.add_argument("--payment", type=float) p.add_argument("--fee", type=float, default=0.0) p.add_argument("--schedule") p.add_argument("--yearly", action="store_true") p.add_argument("--json", action="store_true") try: a = p.parse_args(argv) except SystemExit as e: return 0 if e.code == 0 else 2 try: n = a.months if a.months else round(a.years * 12) if a.principal <= 0 or n <= 0 or a.rate < 0 or a.extra < 0 or a.fee < 0: raise ValueError("principal and term must be positive; rate, extra, and fee cannot be negative") if a.rate > 100: raise ValueError("rate is in percent per year, e.g. 4.2 for 4.2%") lumps = {} for item in a.lump: num, _, amt = item.partition(":") try: if not num.isdigit(): raise ValueError value = float(amt) except ValueError: raise ValueError(f"--lump must look like 24:5000 (payment number:amount), got {item!r}") from None lumps[int(num)] = lumps.get(int(num), 0.0) + value if a.start: try: y, m = a.start.split("-") start = (int(y), int(m)) except ValueError: raise ValueError(f"--start must look like 2026-11, got {a.start!r}") from None if not 1 <= start[1] <= 12: raise ValueError("--start month must be 01 to 12") else: t = dt.date.today() start = add_months((t.year, t.month), 1) base_pay, base_rows = schedule(a.principal, a.rate, n, payment=None) scenario = bool(a.extra or lumps or a.payment) pay, rows = schedule(a.principal, a.rate, n, payment=a.payment, extra=a.extra, extra_from=a.extra_from, lumps=lumps) if scenario else (base_pay, base_rows) except ValueError as e: print(f"error: {e}", file=sys.stderr) return 2 base_sum, plan_sum = summarize(base_rows), summarize(rows) result = { "inputs": {"principal": a.principal, "annual_rate_pct": a.rate, "term_months": n, "first_payment": fmt_ym(start), "extra_monthly": a.extra, "extra_from_payment": a.extra_from, "lumps": lumps, "fixed_payment": a.payment, "fees": a.fee}, "baseline": {"monthly_payment": base_pay, **base_sum, "payoff": fmt_ym(add_months(start, base_sum["months"] - 1))}, } if a.fee: result["baseline"]["apr_estimate_pct"] = apr_estimate(a.principal, a.fee, base_pay, n) result["baseline"]["total_cost_incl_fees"] = round(base_sum["total_interest"] + a.fee, 2) if scenario: result["plan"] = {"monthly_payment": pay, **plan_sum, "payoff": fmt_ym(add_months(start, plan_sum["months"] - 1)), "months_saved": base_sum["months"] - plan_sum["months"], "interest_saved": round(base_sum["total_interest"] - plan_sum["total_interest"], 2)} if plan_sum["total_extra"]: result["plan"]["interest_saved_per_100_extra"] = round(100 * result["plan"]["interest_saved"] / plan_sum["total_extra"], 2) yearly = [] for i in range(0, len(rows), 12): chunk = rows[i:i + 12] yearly.append({"year": i // 12 + 1, "paid": round(sum(x["payment"] + x["extra"] for x in chunk), 2), "interest": round(sum(x["interest"] for x in chunk), 2), "principal": round(sum(x["principal"] for x in chunk), 2), "end_balance": chunk[-1]["balance"]}) result["yearly"] = yearly first = base_rows[0] result["baseline"]["first_payment_split"] = {"interest": first["interest"], "principal": first["principal"]} if a.schedule: try: with open(a.schedule, "w", newline="", encoding="utf-8") as f: w = csv.writer(f) w.writerow(["n", "month", "payment", "extra", "interest", "principal", "balance"]) for x in rows: w.writerow([x["n"], fmt_ym(add_months(start, x["n"] - 1)), f"{x['payment']:.2f}", f"{x['extra']:.2f}", f"{x['interest']:.2f}", f"{x['principal']:.2f}", f"{x['balance']:.2f}"]) except OSError as e: print(f"error: cannot write schedule {a.schedule}: {e.strerror}", file=sys.stderr) return 2 if a.json: print(json.dumps(result, indent=2)) return 0 b = result["baseline"] print(f"Loan {a.principal:,.2f} at {a.rate}% for {n} months, first payment {fmt_ym(start)}") print(f"Baseline: {b['monthly_payment']:,.2f} per month, interest {b['total_interest']:,.2f}, " f"total paid {b['total_paid']:,.2f}, paid off {b['payoff']} ({b['months']} payments)") if a.fee: print(f" With fees {a.fee:,.2f}: total cost of borrowing {b['total_cost_incl_fees']:,.2f}, APR about {b['apr_estimate_pct']}%") print(f" First payment: {first['interest']:,.2f} interest, {first['principal']:,.2f} principal") if scenario: pl = result["plan"] what = [] if a.payment: what.append(f"fixed payment {a.payment:,.2f}") if a.extra: what.append(f"+{a.extra:,.2f}/month" + (f" from payment {a.extra_from}" if a.extra_from > 1 else "")) if lumps: what.append("lump sums " + ", ".join(f"{v:,.2f} in payment {k}" for k, v in sorted(lumps.items()))) print(f"Plan ({'; '.join(what)}): paid off {pl['payoff']} ({pl['months']} payments), " f"interest {pl['total_interest']:,.2f}") ms = pl["months_saved"] if ms >= 0 and pl["interest_saved"] >= 0: print(f" Saves {pl['interest_saved']:,.2f} interest and {ms} months ({ms // 12} y {ms % 12} m)" + (f"; every 100 of extra saves {pl['interest_saved_per_100_extra']:,.2f}" if pl.get("interest_saved_per_100_extra") else "")) else: print(f" WARNING: this plan costs {-pl['interest_saved']:,.2f} MORE interest and takes {-ms} months longer than the baseline") if a.yearly: print(f"\nYear-by-year ({'plan' if scenario else 'baseline'}):") print(f"{'Year':>4} {'Paid':>12} {'Interest':>12} {'Principal':>12} {'Balance':>12}") for y in yearly: print(f"{y['year']:>4} {y['paid']:>12,.2f} {y['interest']:>12,.2f} {y['principal']:>12,.2f} {y['end_balance']:>12,.2f}") if a.schedule: print(f"\nSchedule written to {a.schedule} ({len(rows)} rows)") return 0 if __name__ == "__main__": sys.exit(main(sys.argv[1:]))
Reviews Dockerfiles for security, image size, build cache, and runtime reliability: secrets baked into layers, running as root, unpinned base images, curl piped to a shell, ADD misuse, leftover apt and pip caches, cache-busting dependency installs, shell-form CMD, and missing multi-stage builds or health checks, then writes a corrected Dockerfile and .dockerignore. Use when a user shares a Dockerfile and asks "review my Dockerfile", "why is my image so big?", or "is this container secure?".
---
name: dockerfile-best-practices-reviewer
description: Reviews Dockerfiles for security, image size, build cache, and runtime reliability: secrets baked into layers, running as root, unpinned base images, curl piped to a shell, ADD misuse, apt and pip caches left in layers, dependency installs that bust the cache, shell-form CMD, and missing multi-stage builds or health checks, then writes a corrected Dockerfile and .dockerignore. Use when a user shares a Dockerfile and asks "review my Dockerfile", "why is my image so big?", "is this container secure?", or "why does every build reinstall dependencies?". Includes a tested stdlib Python linter.
---
# Dockerfile Best Practices Reviewer
You help developers ship container images that are small, fast to rebuild, and safe to run. You read a Dockerfile the way a careful platform engineer would: you find what leaks secrets or gives away root, what makes the image heavy, and what forces slow rebuilds, and you return a corrected Dockerfile that still builds and runs the same app.
## Files in this skill
- `scripts/lint_dockerfile.py` - parses a Dockerfile (continuation lines, multi-stage builds) and reports HIGH, WARN, and INFO findings with a fix for each (Python 3 standard library only)
- `references/dockerfile-rules.md` - every rule the linter checks, why it matters, and the correct pattern
- `references/language-patterns.md` - proven multi-stage patterns for Node.js, Python, Go, and Java, plus a starter .dockerignore
- `templates/dockerfile-review.md` - the review report to return
- `examples/example-node-api-review.md` - a worked review of a Node.js API Dockerfile
## Workflow
### 1. Collect the facts
Ask, or assume and say so:
- The Dockerfile, and the .dockerignore if there is one.
- Language and framework, how the app is built and started, and the port.
- Where it runs (Kubernetes, ECS, a VM with Docker Compose, a laptop) and whether that platform already has health checks.
- Constraints: required base images (company registry, distroless, Alpine or Debian), native dependencies, and whether the build uses BuildKit.
### 2. Run the linter
```bash
python3 scripts/lint_dockerfile.py Dockerfile
python3 scripts/lint_dockerfile.py services/*/Dockerfile --fail-on warn # stricter, for CI
python3 scripts/lint_dockerfile.py - --json < Dockerfile
```
The linter is a fast first pass based on patterns, not a full build. Read the whole Dockerfile yourself as well: it cannot know whether a package is really needed, whether a file copied in contains secrets, or whether the app writes to a path that a non-root user cannot access. If you cannot run the script, apply `references/dockerfile-rules.md` by hand.
### 3. Fix in this order
1. Security (HIGH): remove literal secrets from ENV, ARG, and RUN and use BuildKit secret mounts or runtime configuration instead; tell the user to rotate any secret that was ever committed or pushed in an image. Add a non-root USER. Replace curl-pipe-shell and chmod 777.
2. Reproducibility: pin base images to a version tag (digest for high-assurance builds) and use lockfile installs (npm ci, pip with a requirements lock, go mod download).
3. Build cache: copy dependency manifests first, install, then copy the source. Combine update and install in one RUN.
4. Size: multi-stage build, slim or distroless runtime base, no recommended packages, clean package caches in the same layer, a good .dockerignore.
5. Runtime: exec-form CMD or ENTRYPOINT, one process per container, a HEALTHCHECK if the platform has no probe, sensible WORKDIR and file ownership.
Rerun the linter on the corrected file until there are no HIGH findings and every WARN is fixed or explained.
### 4. Deliver
Fill in `templates/dockerfile-review.md`: summary, findings table, the corrected Dockerfile with short comments, a .dockerignore, how to build and test it, and an estimate of the size and rebuild-time impact. Use patterns from `references/language-patterns.md` and follow the style of `examples/example-node-api-review.md`.
## Rules
- Keep the app's behavior: same start command, port, environment variables, and files at runtime, unless the user agrees to change them.
- Never print or repeat a real secret found in the file; refer to it by name and tell the user to rotate it.
- Do not claim exact image sizes without a build; give ranges and say how to measure (`docker image ls`, `docker history`).
- Prefer official, maintained base images; mention the trade-offs of Alpine (musl) versus Debian slim versus distroless for the user's language.
- Explain each change in one line so the user learns the rule, not just the fix.
FILE:references/dockerfile-rules.md
# Dockerfile rules checked by the linter
Each rule: what is detected, why it matters, and the pattern to use instead.
## HIGH (security)
### secret-in-env / secret-in-run
Detected: ENV or ARG with a name like PASSWORD, SECRET, TOKEN, API_KEY, ACCESS_KEY, PRIVATE_KEY, CREDENTIALS and a literal value; `--password=...` or `user:pass@` URLs in RUN; ARG names that look like secrets (WARN).
Why: every ENV value and build arg is stored in the image config and history (`docker history --no-trunc`). Anyone who can pull the image can read it. Deleting it in a later layer does not remove it.
Use instead:
```dockerfile
# syntax=docker/dockerfile:1
RUN --mount=type=secret,id=npm_token \
NPM_TOKEN="$(cat /run/secrets/npm_token)" npm ci
```
Build with `docker build --secret id=npm_token,src=$HOME/.npm_token .`. Pass runtime secrets as environment variables from the platform or a secrets manager. Rotate any secret that was ever in an image.
### root-user
Detected: no USER in the final stage, or USER root / 0.
Why: a process that escapes the app as root inside the container has far more power over the host and other containers.
Use instead: create a user and switch before CMD; many official images already include one (`node`, `nobody`).
```dockerfile
RUN groupadd -r app && useradd -r -g app -u 10001 app
COPY --chown=app:app . /app
USER app
```
### curl-pipe-shell
Detected: `curl ... | sh` or `wget ... | bash`.
Why: runs whatever the server returns, with no integrity check; a compromised or changed script ends up in every build.
Use instead: download, verify a checksum, then run.
```dockerfile
RUN curl -fsSLo /tmp/install.sh https://example.com/install.sh \
&& echo "<sha256> /tmp/install.sh" | sha256sum -c - \
&& sh /tmp/install.sh && rm /tmp/install.sh
```
### chmod-777
Detected: `chmod 777` or `chmod -R 777`.
Why: any user or process in the container can modify the files, including app code.
Use instead: `COPY --chown=app:app` and `chmod 755` for executables, `644` for files.
## WARN
| Rule | Detected | Why | Fix |
|---|---|---|---|
| unpinned-base | FROM without tag or with :latest | builds change silently; hard to reproduce or roll back | `FROM node:20.11-bookworm-slim` or pin a digest `@sha256:...` |
| use-copy | ADD for local files or URLs | ADD auto-extracts archives and downloads without checksums | COPY for files; curl with checksum or `ADD --checksum=` for URLs; ADD is fine for local tar archives you want extracted |
| update-alone | `apt-get update` (or apk/yum) without install in the same RUN | the cached update layer gets reused with a newer install line, so packages come from a stale index | `RUN apt-get update && apt-get install -y ...` |
| cache-bust | dependency install after `COPY . .` | any source change invalidates the install layer, so every build reinstalls everything | copy manifests first, install, then copy the source |
| sudo | sudo in RUN | not needed (build runs as root until USER), adds a setuid binary | run the step before USER |
| ssh-port | EXPOSE 22 | containers should not run SSH daemons | `docker exec`, `kubectl exec` |
| multiple-cmd | more than one CMD or ENTRYPOINT in a stage | only the last one counts; usually a mistake | keep one |
## INFO
| Rule | Fix |
|---|---|
| apt-recommends | `apt-get install -y --no-install-recommends ...` |
| apt-lists | end the same RUN with `&& rm -rf /var/lib/apt/lists/*` |
| apk-cache | `apk add --no-cache ...` |
| pip-cache | `pip install --no-cache-dir ...` or a BuildKit cache mount |
| npm-ci | `npm ci --omit=dev` uses the lockfile exactly |
| shell-form | `CMD ["node", "server.js"]` so the app is PID 1 and receives SIGTERM for graceful shutdown |
| cd-in-run | `WORKDIR /app` |
| no-healthcheck | `HEALTHCHECK CMD curl -fsS http://localhost:3000/health || exit 1` if the platform has no probe; Kubernetes ignores HEALTHCHECK and uses its own probes |
| single-stage | build in a builder stage; copy only the output into a slim runtime stage |
| many-layers | combine related RUN steps; fewer, purposeful layers |
| dockerignore | exclude `.git`, `node_modules`, `.env*`, build output, logs, and local secrets |
## What the linter cannot see
- Whether copied files contain secrets (check .dockerignore and the repo).
- Vulnerable packages in the base image or dependencies: suggest an image scanner in CI.
- Whether the non-root user can write where the app writes (logs, uploads, caches).
- Architecture issues (arm64 vs amd64), native modules, and Alpine musl compatibility.
FILE:references/language-patterns.md
# Multi-stage patterns by language
Adapt versions to the project. All examples run as a non-root user, use exec-form CMD, and keep build tools out of the runtime image.
## Node.js (npm)
```dockerfile
# syntax=docker/dockerfile:1
FROM node:20.11-bookworm-slim AS deps
WORKDIR /app
COPY package.json package-lock.json ./
RUN npm ci --omit=dev
FROM node:20.11-bookworm-slim AS build
WORKDIR /app
COPY package.json package-lock.json ./
RUN npm ci
COPY . .
RUN npm run build
FROM node:20.11-bookworm-slim
ENV NODE_ENV=production
WORKDIR /app
COPY --from=deps /app/node_modules ./node_modules
COPY --from=build --chown=node:node /app/dist ./dist
USER node
EXPOSE 3000
CMD ["node", "dist/server.js"]
```
Skip the build stage if there is no compile step. Use `node:<version>-alpine` only if native modules work with musl.
## Python (pip)
```dockerfile
FROM python:3.12-slim AS build
WORKDIR /app
RUN python -m venv /venv
ENV PATH="/venv/bin:$PATH"
COPY requirements.txt .
RUN pip install --no-cache-dir -r requirements.txt
FROM python:3.12-slim
ENV PATH="/venv/bin:$PATH" PYTHONDONTWRITEBYTECODE=1 PYTHONUNBUFFERED=1
WORKDIR /app
COPY --from=build /venv /venv
RUN useradd -r -u 10001 app
COPY --chown=app:app . .
USER app
EXPOSE 8000
CMD ["gunicorn", "-b", "0.0.0.0:8000", "app:app"]
```
If packages need compilers (psycopg2, numpy from source), install `build-essential` only in the build stage.
## Go
```dockerfile
FROM golang:1.22 AS build
WORKDIR /src
COPY go.mod go.sum ./
RUN go mod download
COPY . .
RUN CGO_ENABLED=0 go build -trimpath -ldflags="-s -w" -o /out/app ./cmd/app
FROM gcr.io/distroless/static-debian12:nonroot
COPY --from=build /out/app /app
USER nonroot
ENTRYPOINT ["/app"]
```
A static Go binary on distroless or scratch gives images of a few MB to tens of MB.
## Java (Maven)
```dockerfile
FROM maven:3.9-eclipse-temurin-21 AS build
WORKDIR /src
COPY pom.xml .
RUN mvn -q dependency:go-offline
COPY src ./src
RUN mvn -q package -DskipTests
FROM eclipse-temurin:21-jre
RUN useradd -r -u 10001 app
WORKDIR /app
COPY --from=build --chown=app:app /src/target/app.jar app.jar
USER app
EXPOSE 8080
ENTRYPOINT ["java", "-XX:MaxRAMPercentage=75", "-jar", "app.jar"]
```
## Starter .dockerignore
```
.git
.gitignore
.dockerignore
Dockerfile*
node_modules
dist
build
target
__pycache__
*.pyc
.venv
.env
.env.*
*.log
coverage
.idea
.vscode
secrets/
*.pem
*.key
```
## Measuring the result
- `docker build -t app:review .` then `docker image ls app:review` for size.
- `docker history app:review` to see which layers are large.
- Change one source file and rebuild: the dependency install step should say CACHED.
FILE:templates/dockerfile-review.md
# Dockerfile review: service_name
**App:** language_and_framework, started with start_command, port port
**Runs on:** platform
**Linter:** HIGH high / WARN warn / INFO info before; HIGH 0 / WARN warn_after / INFO info_after after
## Summary
Two or three sentences: the biggest security problem, the biggest size or speed problem, and the expected improvement.
## Findings
| # | Level | Line | Problem | Fix |
|---|---|---|---|---|
| 1 | HIGH | | | |
## Action needed outside the Dockerfile
- Rotate: (names of any secrets that were in the file or image, never their values)
- Platform: health probes, read-only filesystem, resource limits
## Corrected Dockerfile
```dockerfile
(full corrected Dockerfile with one-line comments on changed parts)
```
## .dockerignore
```
(entries)
```
## Build and verify
```bash
docker build -t image:review .
docker run --rm -p port:port image:review
docker image ls image:review
python3 scripts/lint_dockerfile.py Dockerfile
```
## Expected impact (estimate)
- Image size: from about ... to about ...
- Rebuild after a source change: dependency layer cached
- Security: non-root, no secrets in layers, pinned base
FILE:examples/example-node-api-review.md
# Example: review of a Node.js API Dockerfile
## User request
"Our orders API image is 1.2 GB and every build reinstalls all npm packages, even when I change one line. Can you review the Dockerfile? It runs on Kubernetes, Express app, `node server.js`, port 3000. No native modules as far as I know."
## Input Dockerfile
```dockerfile
FROM node
ENV NODE_ENV=production
ENV DB_PASSWORD=supersecret123
RUN apt-get update
RUN apt-get install -y curl git build-essential
WORKDIR /app
COPY . .
RUN npm install
RUN curl -fsSL https://example.com/install-agent.sh | bash
RUN chmod -R 777 /app
ADD config.json /app/config.json
EXPOSE 3000 22
CMD node server.js
```
## Linter run
```
$ python3 scripts/lint_dockerfile.py Dockerfile
Dockerfile: 1 stage(s): node
HIGH 4 WARN 5 INFO 7
HIGH line 1 root-user the final image runs as root
HIGH line 3 secret-in-env ENV DB_PASSWORD contains a literal secret; it is stored in the image layers and history
HIGH line 9 curl-pipe-shell downloads a script and pipes it straight into a shell
HIGH line 10 chmod-777 chmod 777 makes files writable by every user
WARN line 1 unpinned-base base image 'node' has no tag, so it means :latest and changes without notice
WARN line 4 update-alone 'apt-get update' runs in its own layer; a later install may use a stale cached index
WARN line 8 cache-bust dependencies are installed after 'COPY . .' (line 7), so any source change reinstalls them
WARN line 11 use-copy ADD is used for local files; it also auto-extracts archives and fetches URLs
WARN line 12 ssh-port port 22 (SSH) is exposed; containers should not run an SSH server
INFO line 1 no-healthcheck no HEALTHCHECK in the final stage
INFO line 1 single-stage build tools are installed in the only stage, so they ship in the final image
INFO line 5 apt-recommends apt installs recommended packages too
INFO line 5 apt-lists apt package lists are left in the layer (tens of MB)
INFO line 7 dockerignore COPY of the whole build context
INFO line 8 npm-ci npm install can change the lockfile and versions
INFO line 13 shell-form shell-form CMD runs under /bin/sh -c, so the app is not PID 1 and may miss SIGTERM
```
## Review (filled template, shortened)
**Linter:** HIGH 4 / WARN 5 / INFO 7 before; HIGH 0 / WARN 0 / INFO 0 after
**Summary:** The image ships a database password in its layers and runs as root, and it pipes a remote script into bash. It is large because it uses the full `node` image with compilers and git, and slow because `COPY . .` comes before `npm install`. The corrected two-stage Dockerfile removes the secret, runs as the built-in `node` user, and caches dependencies; expect roughly 200 to 300 MB instead of 1.2 GB (measure with `docker image ls`).
| # | Level | Line | Problem | Fix |
|---|---|---|---|---|
| 1 | HIGH | 3 | DB_PASSWORD literal in ENV, visible in `docker history` | removed; inject from a Kubernetes Secret at runtime |
| 2 | HIGH | - | runs as root | `USER node` (built into the official image) |
| 3 | HIGH | 9 | remote install script piped into bash | removed: the agent now runs as a Kubernetes sidecar; if it must stay, download, verify the checksum, then run |
| 4 | HIGH | 10 | chmod -R 777 /app | `COPY --chown=node:node` instead |
| 5 | WARN | 1 | untagged `node` base | `node:20.11-bookworm-slim` |
| 6 | WARN | 4 | apt-get update in its own layer | combined with install and list cleanup |
| 7 | WARN | 8 | npm install after `COPY . .` | manifests first, `npm ci --omit=dev` in a deps stage |
| 8 | WARN | 11 | ADD for a local file | COPY |
| 9 | WARN | 12 | EXPOSE 22 | removed; use `kubectl exec` |
| 10 | INFO | 5, 13 | build tools and git in the runtime image, shell-form CMD | dropped (no native modules), exec-form CMD |
**Action needed outside the Dockerfile:** rotate DB_PASSWORD now (it is in every image pushed so far), and add it to the Deployment from a Kubernetes Secret. Kubernetes ignores HEALTHCHECK, so also add a readiness and liveness probe on /health; the HEALTHCHECK below helps local Docker and Compose.
## Corrected Dockerfile
```dockerfile
# syntax=docker/dockerfile:1
FROM node:20.11-bookworm-slim AS deps
WORKDIR /app
COPY package.json package-lock.json ./
RUN npm ci --omit=dev
FROM node:20.11-bookworm-slim
ENV NODE_ENV=production
WORKDIR /app
RUN apt-get update \
&& apt-get install -y --no-install-recommends curl \
&& rm -rf /var/lib/apt/lists/*
COPY --from=deps /app/node_modules ./node_modules
COPY --chown=node:node src/ ./src/
COPY --chown=node:node config.json ./
USER node
EXPOSE 3000
HEALTHCHECK --interval=30s --timeout=3s CMD curl -fsS http://localhost:3000/health || exit 1
CMD ["node", "src/server.js"]
```
## .dockerignore
```
.git
node_modules
.env
.env.*
*.log
coverage
Dockerfile*
```
## Verify
```
$ python3 scripts/lint_dockerfile.py Dockerfile
Dockerfile: 2 stage(s): node:20.11-bookworm-slim AS deps, node:20.11-bookworm-slim
HIGH 0 WARN 0 INFO 0
no problems found
```
Note: the source now lives in `src/`, so the start command became `node src/server.js`; if your entry file is elsewhere, keep your original path. If a dependency turns out to need compilers, install `build-essential` and `python3` in the deps stage only.
FILE:scripts/lint_dockerfile.py
#!/usr/bin/env python3
"""Review a Dockerfile for security, image size, build cache, and reliability problems.
Usage:
python3 lint_dockerfile.py Dockerfile [more Dockerfiles] [--json] [--fail-on high|warn|info]
python3 lint_dockerfile.py - < Dockerfile
Checks (rule ids in brackets):
HIGH secrets in ENV/ARG or in RUN (secret-in-env, secret-in-run), final stage runs as root
(root-user), pipe from curl/wget to a shell (curl-pipe-shell), chmod 777 (chmod-777)
WARN untagged or :latest base image (unpinned-base), ADD for local files or URLs (use-copy),
apt/apk/yum update without install in the same RUN (update-alone), COPY of the whole
context before installing dependencies (cache-bust), sudo (sudo), EXPOSE 22 (ssh-port),
more than one CMD or ENTRYPOINT in a stage (multiple-cmd)
INFO apt-get without --no-install-recommends or without cleaning lists, pip without
--no-cache-dir, npm install instead of npm ci, shell-form CMD/ENTRYPOINT, cd in RUN
instead of WORKDIR, no HEALTHCHECK, single-stage build with build tools, many RUN layers,
COPY of the whole context (check .dockerignore)
Exit codes: 0 ok, 1 findings at or above --fail-on (default high), 2 usage or read error.
Python 3 standard library only.
"""
import json
import re
import shlex
import sys
LEVELS = {"info": 0, "warn": 1, "high": 2}
SECRET_KEY = re.compile(r"(PASSWORD|PASSWD|SECRET|TOKEN|API_?KEY|ACCESS_?KEY|PRIVATE_?KEY|CREDENTIALS?)", re.I)
BUILD_TOOLS = re.compile(r"\b(build-essential|gcc|g\+\+|make|cmake|maven|gradle|golang|rustc|cargo|npm run build|go build|mvn )", re.I)
def logical_lines(text):
"""Join continuation lines; return list of (line_no, instruction, args). Skips comments and parser directives."""
out, buf, start = [], "", None
escape = "\\"
for i, raw in enumerate(text.replace("\r\n", "\n").split("\n"), 1):
m = re.match(r"^#\s*escape\s*=\s*(\S)", raw)
if m and not out and not buf:
escape = m.group(1)
continue
s = raw.strip()
if not buf and (not s or s.startswith("#")):
continue
if buf and s.startswith("#"):
continue
if start is None:
start = i
if s.endswith(escape):
buf += s[:-1] + " "
continue
buf += s
parts = buf.split(None, 1)
out.append((start, parts[0].upper(), parts[1] if len(parts) > 1 else ""))
buf, start = "", None
if buf:
parts = buf.split(None, 1)
out.append((start, parts[0].upper(), parts[1] if len(parts) > 1 else ""))
return out
def review(text):
lines = logical_lines(text)
findings = []
def add(level, rule, line, msg, fix):
findings.append({"level": level, "rule": rule, "line": line, "message": msg, "fix": fix})
if not any(ins == "FROM" for _, ins, _ in lines):
raise ValueError("no FROM instruction found (is this a Dockerfile?)")
stages, cur = [], None
for ln, ins, args in lines:
if ins == "FROM":
m = re.match(r"(?:--platform=\S+\s+)?(\S+)(?:\s+AS\s+(\S+))?", args, re.I)
cur = {"line": ln, "image": m.group(1) if m else args, "name": m.group(2) if m else None, "ins": []}
stages.append(cur)
elif cur is not None:
cur["ins"].append((ln, ins, args))
names = {s["name"].lower() for s in stages if s["name"]}
arg_names = {a.split("=")[0].strip() for _, i, a in lines if i == "ARG"}
for s in stages:
img = s["image"]
if img.lower() in names or img == "scratch" or img.startswith("$"):
pass
elif "@sha256:" in img:
pass
elif ":" not in img.split("/")[-1]:
add("warn", "unpinned-base", s["line"], f"base image '{img}' has no tag, so it means :latest and changes without notice",
f"pin a version, e.g. {img}:<version>-slim, or a digest")
elif img.endswith(":latest"):
add("warn", "unpinned-base", s["line"], f"base image '{img}' uses :latest", "pin a specific version tag or digest")
final = stages[-1]
run_count = 0
for idx, s in enumerate(stages):
cmds = [x for x in s["ins"] if x[1] in ("CMD", "ENTRYPOINT")]
for kind in ("CMD", "ENTRYPOINT"):
k = [x for x in cmds if x[1] == kind]
if len(k) > 1:
add("warn", "multiple-cmd", k[-1][0], f"{len(k)} {kind} instructions in one stage; only the last one counts",
f"keep a single {kind}")
copied_all = None
for ln, ins, args in s["ins"]:
low = args.lower()
if ins in ("ENV", "ARG"):
pairs = re.findall(r"([A-Za-z_][A-Za-z0-9_]*)(?:(?:=|\s+)(\"[^\"]*\"|'[^']*'|[^\s=]\S*))?", args) if ins == "ARG" or "=" in args else [tuple((args.split(None, 1) + [""])[:2])]
for key, val in pairs:
if SECRET_KEY.search(key) and val and not val.startswith("$") and val.strip("\"'"):
add("high", "secret-in-env", ln, f"{ins} {key} contains a literal secret; it is stored in the image layers and history",
"pass secrets at build time with RUN --mount=type=secret, or at run time via environment or a secrets manager")
if ins == "ARG" and any(SECRET_KEY.search(k) for k, _ in pairs) and not any(v for k, v in pairs if SECRET_KEY.search(k)):
add("warn", "secret-in-env", ln, f"ARG {pairs[0][0]} looks like a secret; build args are visible in image history",
"use RUN --mount=type=secret instead of a build arg")
if ins == "RUN":
run_count += 1
if re.search(r"(curl|wget)\b[^|;&]*\|\s*(sudo\s+)?(ba|z|da)?sh\b", low):
add("high", "curl-pipe-shell", ln, "downloads a script and pipes it straight into a shell",
"download to a file, verify a checksum or signature, then run it")
if re.search(r"chmod\s+(-r\s+)?0?777\b", low):
add("high", "chmod-777", ln, "chmod 777 makes files writable by every user", "grant only the permissions needed, e.g. chmod 755 or chown to the app user")
if re.search(r"(--password[= ]\S+|://[^/\s:@]+:[^@\s$]+@)", args) or re.search(r"\b(token|api[_-]?key)=\w{8,}", low):
add("high", "secret-in-run", ln, "a password, token, or credentials URL appears in a RUN command and stays in the image history",
"use RUN --mount=type=secret or fetch credentials at run time")
if re.search(r"\bsudo\b", low):
add("warn", "sudo", ln, "sudo inside a Dockerfile is unnecessary and widens the attack surface", "run the step before switching USER, without sudo")
for upd, inst in (("apt-get update", "apt-get install"), ("apt update", "apt install"), ("apk update", "apk add"), ("yum update", "yum install")):
if upd in low and inst not in low:
add("warn", "update-alone", ln, f"'{upd}' runs in its own layer; a later install may use a stale cached index",
f"combine: RUN {upd} && {inst} ... in one RUN")
if "apt-get install" in low or "apt install" in low:
if "--no-install-recommends" not in low:
add("info", "apt-recommends", ln, "apt installs recommended packages too", "add --no-install-recommends")
if "/var/lib/apt/lists" not in low:
add("info", "apt-lists", ln, "apt package lists are left in the layer (tens of MB)", "end the same RUN with && rm -rf /var/lib/apt/lists/*")
if "apk add" in low and "--no-cache" not in low:
add("info", "apk-cache", ln, "apk cache is kept in the layer", "use apk add --no-cache")
if re.search(r"\bpip3?\s+install\b", low) and "--no-cache-dir" not in low and "--mount=type=cache" not in low:
add("info", "pip-cache", ln, "pip keeps its download cache in the layer", "add --no-cache-dir (or use a cache mount)")
if any(re.fullmatch(r"\s*npm\s+(install|i)(\s+--?(?!g\b|global\b)[\w-]+(=\S+)?)*\s*", seg) for seg in re.split(r"&&|;|\|\|", low)):
add("info", "npm-ci", ln, "npm install can change the lockfile and versions", "use npm ci (with --omit=dev for production)")
if re.match(r"cd\s+\S+\s*&&", low) or re.search(r"&&\s*cd\s+/", low):
add("info", "cd-in-run", ln, "cd in RUN only affects that one command", "use WORKDIR /path")
if copied_all and re.search(r"\b(npm (ci|install)|pip3? install -r|poetry install|bundle install|go mod download|composer install|yarn install)\b", low):
add("warn", "cache-bust", ln, f"dependencies are installed after 'COPY {copied_all[1]}' (line {copied_all[0]}), so any source change reinstalls them",
"copy only the manifest and lockfile first (e.g. COPY package*.json ./), install, then COPY the rest")
copied_all = None
if ins == "ADD":
src = args.split()[0] if args.split() else ""
if re.match(r"https?://", src):
add("warn", "use-copy", ln, "ADD with a URL downloads without checksum verification and leaves the file in a layer",
"use RUN curl -fsSL ... with a checksum check, or ADD --checksum=sha256:...")
elif not re.search(r"\.(tar|tar\.gz|tgz|tar\.xz|tar\.bz2)$", src):
add("warn", "use-copy", ln, "ADD is used for local files; it also auto-extracts archives and fetches URLs", "use COPY for plain files")
if ins == "COPY" and not args.startswith("--from"):
parts = [p for p in args.split() if not p.startswith("--")]
if parts and parts[0] in (".", "./", "*"):
copied_all = (ln, " ".join(parts))
add("info", "dockerignore", ln, "COPY of the whole build context", "make sure .dockerignore excludes .git, node_modules, .env, build output, and secrets")
if ins == "EXPOSE" and re.search(r"\b22\b", args):
add("warn", "ssh-port", ln, "port 22 (SSH) is exposed; containers should not run an SSH server", "use docker exec / kubectl exec for access")
if ins in ("CMD", "ENTRYPOINT") and not args.strip().startswith("["):
add("info", "shell-form", ln, f"shell-form {ins} runs under /bin/sh -c, so the app is not PID 1 and may miss SIGTERM",
f'use exec form: {ins} ["executable", "arg"]')
users = [(ln, a.strip()) for ln, i, a in final["ins"] if i == "USER"]
if not users or users[-1][1].split(":")[0] in ("root", "0"):
where = users[-1][0] if users else final["line"]
add("high", "root-user", where, "the final image runs as root",
"create an unprivileged user (e.g. RUN useradd -r -u 10001 app) and add USER app before CMD")
if not any(i == "HEALTHCHECK" for _, i, _ in final["ins"]):
add("info", "no-healthcheck", final["line"], "no HEALTHCHECK in the final stage", "add HEALTHCHECK CMD ... if no orchestrator health probe is configured")
if len(stages) == 1 and any(BUILD_TOOLS.search(a) for _, i, a in final["ins"] if i == "RUN"):
add("info", "single-stage", final["line"], "build tools are installed in the only stage, so they ship in the final image",
"use a multi-stage build: compile in a builder stage, COPY --from=builder only the output")
if run_count > 10:
add("info", "many-layers", final["line"], f"{run_count} RUN instructions; related steps can be combined", "group related commands with && in fewer RUN steps")
return {"stages": [{"line": s["line"], "image": s["image"], "name": s["name"]} for s in stages], "findings": findings}
def main(argv):
as_json, fail_on, files = False, "high", []
it = iter(argv)
for a in it:
if a == "--json":
as_json = True
elif a == "--fail-on":
fail_on = next(it, "").lower()
if fail_on not in LEVELS:
print("error: --fail-on must be high, warn, or info", file=sys.stderr)
return 2
elif a in ("-h", "--help"):
print(__doc__)
return 0
elif a.startswith("--"):
print(f"error: unknown option {a}", file=sys.stderr)
return 2
else:
files.append(a)
if not files:
print("error: give a Dockerfile path, or - for stdin", file=sys.stderr)
return 2
reports, worst = [], -1
for f in files:
try:
text = sys.stdin.read() if f == "-" else open(f, encoding="utf-8").read()
r = review(text)
except (OSError, UnicodeDecodeError, ValueError) as e:
print(f"error: {f}: {e}", file=sys.stderr)
return 2
r["file"] = f
r["findings"].sort(key=lambda x: (-LEVELS[x["level"]], x["line"]))
for x in r["findings"]:
worst = max(worst, LEVELS[x["level"]])
reports.append(r)
if as_json:
print(json.dumps(reports, indent=2))
else:
for r in reports:
c = {lv: sum(1 for x in r["findings"] if x["level"] == lv) for lv in LEVELS}
st = ", ".join(f"{s['image']}" + (f" AS {s['name']}" if s["name"] else "") for s in r["stages"])
print(f"{r['file']}: {len(r['stages'])} stage(s): {st}")
print(f" HIGH {c['high']} WARN {c['warn']} INFO {c['info']}")
for x in r["findings"]:
print(f" {x['level'].upper():<4} line {x['line']:<3} {x['rule']:<15} {x['message']}")
print(f" fix: {x['fix']}")
if not r["findings"]:
print(" no problems found")
return 1 if worst >= LEVELS[fail_on] else 0
if __name__ == "__main__":
sys.exit(main(sys.argv[1:]))Explains Python regular expressions token by token, tests them on must-match and must-not-match samples with spans and named groups, and flags risky patterns: nested quantifiers that backtrack catastrophically, [A-z], unescaped dots, empty matches, and missing anchors, then proposes a safer rewrite. Use when a user asks "what does this regex do?", "why doesn't it match?", or "is this regex safe?". Includes a tested stdlib Python checker.
--- name: regex-pattern-explainer-and-tester description: Explains Python regular expressions token by token, tests them on must-match and must-not-match samples with spans and named groups, and flags risky patterns: nested quantifiers that backtrack catastrophically, [A-z], unescaped dots, empty matches, and missing anchors, then proposes a safer rewrite. Use when a user asks "what does this regex do?", "why doesn't it match?", or "is this regex safe?". Includes a tested stdlib Python checker. --- # Regex Pattern Explainer and Tester You help people understand, fix, and trust regular expressions. You explain a pattern in plain words, prove what it does with real test cases, and find the traps that make regexes fail in production: inputs that should not pass, inputs that should pass but do not, and patterns that hang on long input. You return a corrected pattern with a test table, not just an opinion. ## Files in this skill - `scripts/regex_check.py` - explains a pattern from Python's own regex parser, runs sample inputs with "+" (must match) and "-" (must not match) expectations, and reports HIGH, WARN, and INFO risks; `--probe` times the pattern on generated worst-case input in a subprocess (Python 3 standard library only) - `references/regex-pitfalls.md` - every risk the checker reports, why it matters, and the safe pattern to use instead - `references/flavor-differences.md` - how Python `re` differs from JavaScript, PCRE, Go RE2, .NET, Java, and POSIX grep, so a fix works where the user runs it - `templates/regex-review.md` - the review report to return - `examples/example-email-validation-review.md` - a worked review of an email check that hangs on long input ## Workflow ### 1. Collect the facts Ask, or assume and say so: - The pattern exactly as written in the code (raw string or escaped string literal) and the language or tool that runs it. - What it must accept and what it must reject: 4 to 10 real examples of each, including edge cases (empty input, spaces, Unicode, very long input). - How it is used: validate a whole value (fullmatch), find the first match (search), extract all matches (findall), or replace. - Where the input comes from. Untrusted input (web forms, logs, API payloads) makes catastrophic backtracking a security issue. ### 2. Run the checker ```bash python3 scripts/regex_check.py 'PATTERN' -t 'sample one' -t 'sample two' python3 scripts/regex_check.py 'PATTERN' --cases cases.txt --mode fullmatch --flags a python3 scripts/regex_check.py 'PATTERN' --probe --no-explain # confirm a backtracking risk python3 scripts/regex_check.py 'PATTERN' --cases - --json < cases.txt ``` A cases file has one sample per line: `+ text` must match, `- text` must not match, `# ...` is a comment. Always quote the pattern in single quotes in the shell. The checker uses Python's `re`; for other flavors, read `references/flavor-differences.md` and translate the result. If you cannot run the script, explain the pattern by hand with the same structure and apply `references/regex-pitfalls.md`. ### 3. Explain, then fix 1. Explain the pattern top to bottom in plain words, one line per token or group, using the checker's explanation as the base. 2. Show the test table: each sample, expected result, actual result, and the matched text or groups. 3. Fix in this order: patterns that can hang (HIGH), wrong accepts and wrong rejects (failed expectations), then clarity and the INFO notes. 4. Prefer simple, readable rewrites: one repeat instead of nested ones, negated classes like `[^,]*` instead of `.*`, explicit anchors `\A...\Z` or `fullmatch`, and `re.VERBOSE` with comments for long patterns. 5. Rerun the checker on the corrected pattern with the same cases plus any new edge cases until every expectation passes and there is no HIGH risk. Run `--probe` on the fix if the original had a HIGH risk. ### 4. Deliver Fill in `templates/regex-review.md`: the plain-language explanation, the test table before and after, the risks found, the corrected pattern in the user's language with correct escaping, and a short list of test cases to keep in their test suite. Follow the style of `examples/example-email-validation-review.md`. ## Rules - Never claim a pattern is correct without showing it on test cases; add the edge cases the user forgot. - Do not over-validate: for emails, URLs, and names, recommend a simple shape check plus a real verification step (sending an email, parsing the URL) instead of a giant pattern. - Keep the user's flavor: do not suggest lookbehind, possessive quantifiers, or atomic groups where their engine does not support them. - Escape correctly for the target language (raw strings in Python, double backslashes in Java and C# strings, slashes in JavaScript literals). - Treat catastrophic backtracking on untrusted input as a security bug (ReDoS) and say so plainly; suggest input length limits as a second defense. FILE:references/regex-pitfalls.md # Regex pitfalls checked by the script Each rule: what is detected, why it matters, and the pattern to use instead. Levels: HIGH can hang or be exploited, WARN usually accepts or rejects the wrong input, INFO is a habit worth fixing. ## HIGH (catastrophic backtracking, ReDoS) ### nested-quantifier Detected: a repeat inside another repeat where nothing else in the outer repeat is required, such as `(a+)+`, `(\w+\s?)+`, `([a-z0-9]+\.?)+`, `(\w+|\d+)+`. Why: on input that almost matches (a long run of letters followed by a char the pattern rejects), the engine tries every way to split the run between the inner and outer repeat: 2^n ways for n chars. 30 chars can take minutes. On a web form or API this is a denial-of-service bug. Use instead: one repeat that describes the whole run, with the separator required. ``` (\w+\s?)+$ -> \w+(?:\s\w+)*$ ([a-z0-9]+\.?)+@ -> [a-z0-9]+(?:\.[a-z0-9]+)*@ ``` Python 3.11+ also supports atomic groups `(?>...)` and possessive quantifiers `a++`, which never give back what they matched. ### ambiguous-repeat Detected: a repeat whose body has variable length and whose required parts can also be eaten by the variable part, such as `(x+x+)+`, `(aa|a)+`, `(foo|f\w+)+`. Why: same as above, the iterations can be cut in many ways. Growth may be slower (Fibonacci instead of 2^n) but is still exponential. Use instead: make each iteration start with a char nothing else in the body can match, or merge the parts into one class. Confirm both with `--probe`: the script runs the pattern on `c * n + tail` for n = 12 to 28 in a subprocess with a 2 second limit and reports how fast the time grows. A growth of more than 3x per 4 extra chars, or a timeout, confirms the risk. ## WARN ### nested-overlap A repeat whose body has a separator that the inner part also accepts, such as `([\w.]+\.)+`. Make the inner class exclude the separator: `(?:[\w-]+\.)+`. ### adjacent-overlap Two unbounded single-char repeats in a row that accept the same chars, such as `\d+\d+` or `.*.*`. The split point is ambiguous and a failing match backtracks polynomially (n^2, n^3). Keep one of them or make the boundary explicit. ### range-A-z `[A-z]` covers ASCII 65 to 122, which includes `[ \ ] ^ _` and the backtick. Use `[A-Za-z]`, or `[^\W\d_]` for any Unicode letter. ### unescaped-dot `.` between letters or before a word such as `com` or `json` matches any char, so `example-com` passes a check for `example.com`. Write `\.`. ### matches-empty The pattern accepts the empty string. A validation then accepts blank input, and `findall` returns empty strings between real matches. Use `+` instead of `*` where at least one char is required. ## INFO ### not-anchored `re.search(r"\d+", "abc123xyz")` succeeds. For validating a whole value, use `re.fullmatch` or anchor with `\A ... \Z`. ### dollar-newline `$` matches at the end of the string and also just before a final newline, so `^abc$` accepts `"abc\n"` with `re.search` or `re.match`. Use `\Z` or `re.fullmatch` for strict checks. ### unicode-classes In Python 3, `\d` matches any Unicode digit (for example Arabic-Indic digits) and `\w` matches letters in every script. Good for names and free text; for IDs, card numbers, and codes use `re.ASCII` or `[0-9]`. ### greedy-dotstar and many-dotstar `.*` takes everything and then gives back one char at a time. Three or more `.*` in one pattern make failures slow on long lines. Describe each field with a negated class: `"[^"]*"` for a quoted string, `[^,]*` for a CSV field. ## Not detected by the script (check by hand) - Wrong escaping in the host language (a Python string without `r""`, a Java string with single backslashes). - Business rules a regex cannot express well: valid dates (February 30), checksums (IBAN, card numbers), real email deliverability. - Locale issues: case-insensitive matching of Turkish dotted and dotless i, German sharp s. - Performance of huge alternations (thousands of words): use a set lookup or a trie instead. FILE:references/flavor-differences.md # Regex flavor differences that break fixes The checker runs Python `re`. Before handing a pattern to another engine, check this table. | Feature | Python `re` | JavaScript | PCRE / PHP | Go (RE2) | .NET | Java | POSIX grep -E | |---|---|---|---|---|---|---|---| | Lookahead `(?=...)` | yes | yes | yes | no | yes | yes | no | | Lookbehind `(?<=...)` | fixed width only | yes (ES2018) | fixed or bounded | no | any width | bounded | no | | Named group | `(?P<name>...)` only | `(?<name>...)` | both | `(?P<name>...)` (also `(?<name>...)` in Go 1.22+) | `(?<name>...)` | `(?<name>...)` | no | | Atomic group `(?>...)` | 3.11+ | no | yes | no | yes | yes | no | | Possessive `a++` | 3.11+ | no | yes | no | no | yes | no | | Backreference `\1` | yes | yes | yes | no | yes | yes | yes (basic) | | `\d` matches non-ASCII digits | yes (unless re.ASCII) | no | no (unless UCP) | no | yes | no (unless UNICODE_CHARACTER_CLASS) | locale | | `$` before final newline | yes | no (without m flag) | yes | no | yes | yes | n/a (line based) | | Inline flags `(?i)` | yes, at start or scoped `(?i:...)` | flags after the literal (`/i`); scoped `(?i:...)` only in the newest engines | yes | yes | yes | yes | no | | Catastrophic backtracking | possible | possible | possible (has limits) | impossible (linear time) | possible (timeout option) | possible | engine dependent | | Full match helper | `re.fullmatch` | anchor with `^...$` | anchor | anchor | anchor | `String.matches()` | `-x` flag | ## Translation notes - JavaScript: write the literal as `/pattern/flags`; escape `/` inside it. Use the `u` flag for Unicode and `v` for set operations in newer engines. `\Z` does not exist: use `$` without the `m` flag. - Go RE2: no lookarounds or backreferences, but every pattern runs in linear time, so ReDoS is not a concern there. Rewrite lookarounds as a second check in code. - Java and C#: every backslash doubles in a normal string literal (`"\\d+"`); C# verbatim strings `@"\d+"` and Java text blocks help. - .NET: set `RegexOptions.NonBacktracking` (.NET 7+) or a match timeout for untrusted input. - SQL: PostgreSQL `~` uses Tcl-style advanced regexes (lookahead, lookbehind since 9.6, no named groups); MySQL 8 uses ICU. Test in the database itself. - Shell: `grep -E` has no `\d` on many systems; use `[0-9]`. `grep -P` gives PCRE where available. ## Safe defaults across flavors - `[0-9]` instead of `\d`, `[A-Za-z]` instead of `\w` when the format is ASCII by definition. - Anchors `^` and `$` without multiline mode, plus a separate length check (for example at most 254 chars for an email). - No nested quantifiers on untrusted input, whatever the engine. FILE:templates/regex-review.md # Regex review: purpose **Pattern (original):** `original_pattern` **Engine:** language_and_engine | **Used as:** validate whole value / find first / find all / replace **Checker:** HIGH high / WARN warn / INFO info before; expectations pass_before/total pass ## What it does, in plain words 1. (one line per token or group, top to bottom) ## Test results | # | Input | Expected | Before | After | Notes | |---|---|---|---|---|---| | 1 | | match / no match | | | | ## Risks found | Level | Rule | What it means here | Fix | |---|---|---|---| | HIGH | | | | Probe (if run): input `probe_input`, probe_result ## Corrected pattern ```language (the corrected pattern, escaped for the user's language, with a VERBOSE or commented version if it is longer than about 40 chars) ``` Why each change: (one line per change) ## Keep these tests ``` + (must match) - (must not match) ``` ## Notes - Limits of a regex for this job and the extra check to do in code (length limit, parsing, verification step). - Flavor caveats if the pattern will run outside Python. FILE:examples/example-email-validation-review.md # Example: review of a signup form email check that hangs ## User request "Our signup API sometimes times out and the profiler points at this line. Can you explain the regex and fix it? It runs in Python 3.12 on the email field of a public form." ```python EMAIL = re.compile(r"^([a-zA-Z0-9]+\.?)+@[A-z]+.com$") ``` ## Cases used (cases.txt) The user's 5 cases plus 3 edge cases added during the review (a dash instead of the dot, an underscore in the domain, and 30 letters with no `@`, the shape a bot sent): ``` + ann@example.com + first.last@mail.example.co - bob@@example.com - no-at-sign.example.com - trailing-dot@example. - ann@example-com - ann@exa_mple.com - aaaaaaaaaaaaaaaaaaaaaaaaaaaaaa! ``` ## Checker run (original) ``` $ python3 scripts/regex_check.py '^([a-zA-Z0-9]+\.?)+@[A-z]+.com$' --cases cases.txt --no-explain --probe Samples (search): PASS + 'ann@example.com' -> [0:15] 'ann@example.com' 1='ann' FAIL + 'first.last@mail.example.co' -> expected a match PASS - 'bob@@example.com' PASS - 'no-at-sign.example.com' PASS - 'trailing-dot@example.' FAIL - 'ann@example-com' -> [0:15] 'ann@example-com' 1='ann' (expected no match) FAIL - 'ann@exa_mple.com' -> [0:16] 'ann@exa_mple.com' 1='ann' (expected no match) TIMEOUT - 'aaaaaaaaaaaaaaaaaaaaaaaaaaaaaa!' -> still running after 2 s (catastrophic backtracking) Risks and suggestions: HIGH nested-quantifier a quantifier sits inside another quantifier and nothing else in the outer repeat is required ... probe 'a' * n + '!': n=12: 0.0002 s, n=16: 0.0031 s, n=20: 0.0593 s, n=24: 0.8419 s, n=28: timeout -> TIMEOUT at length 29: exponential backtracking confirmed WARN range-A-z [A-z] also matches [ \ ] ^ _ and the backtick; use [A-Za-z] WARN unescaped-dot '.' before 'com' matches any char (so a '-' or 'x' passes there too); write \. for a literal dot INFO dollar-newline $ also matches before a trailing newline ... Summary: HIGH 1 WARN 2 INFO 1 | expectations 4/8 pass | 1 sample(s) timed out ``` ## Review (filled template, shortened) **Checker:** HIGH 1 / WARN 2 / INFO 1 before; expectations 4/8 pass (one timeout). After: HIGH 0 / WARN 0, expectations 8/8 pass. What it does: start of string; one or more runs of letters or digits, each optionally followed by a dot; an `@`; one or more chars from A to z (including `[ \ ] ^ _` and the backtick); any single char; the text `com`; end of string or before a final newline. | # | Input | Expected | Before | After | Notes | |---|---|---|---|---|---| | 1 | ann@example.com | match | match | match | | | 2 | first.last@mail.example.co | match | no match | match | only `.com` and one domain label were allowed | | 3 | bob@@example.com | no match | no match | no match | | | 4 | no-at-sign.example.com | no match | no match | no match | | | 5 | trailing-dot@example. | no match | no match | no match | | | 6 | ann@example-com | no match | match | no match | added: the unescaped dot accepted `-` | | 7 | ann@exa_mple.com | no match | match | no match | added: `[A-z]` accepted `_` | | 8 | 30 x "a" then "!" | no match, fast | timeout after 2 s | no match, instant | added: the production hang | Risks: the `([a-zA-Z0-9]+\.?)+` group lets a run of letters be split in 2^n ways when the `@` is missing, so a 29-char junk input from a bot blocked a worker for more than 2 seconds; this is a ReDoS bug on a public endpoint. Corrected pattern: ```python EMAIL = re.compile(r""" [A-Za-z0-9._%+-]+ # local part: one run, dots allowed but not repeated in a nested group @ (?:[A-Za-z0-9-]+\.)+ # one or more domain labels, each ending in a literal dot [A-Za-z]{2,} # top-level domain """, re.VERBOSE | re.ASCII) def looks_like_email(value: str) -> bool: return len(value) <= 254 and EMAIL.fullmatch(value) is not None ``` Why each change: one flat repeat for the local part removes the nested quantifier; `\.` makes the dot literal; `[A-Za-z]` replaces `[A-z]`; any top-level domain is accepted instead of only `com`; `fullmatch` replaces `^...$`; a 254-char limit is a second defense; `re.ASCII` keeps `[...]` classes predictable. Recheck: `python3 scripts/regex_check.py '[A-Za-z0-9._%+-]+@(?:[A-Za-z0-9-]+\.)+[A-Za-z]{2,}' --cases cases.txt --mode fullmatch --probe` reports no HIGH or WARN risks (the nested group is reported as INFO nested-separated, a safe shape) and 8/8 expectations pass. Keep these tests: all 8 rows above, plus a 300-char input to check the length limit. Note that a regex only checks the shape; confirm the address by sending a verification email. FILE:scripts/regex_check.py #!/usr/bin/env python3 """Explain a Python regular expression, test it on sample inputs, and warn about risky patterns. Usage: python3 regex_check.py PATTERN [options] python3 regex_check.py PATTERN -t "text one" -t "text two" python3 regex_check.py PATTERN --cases cases.txt --mode fullmatch python3 regex_check.py PATTERN --cases - < cases.txt (read cases from stdin) Options: -t, --text TEXT a sample input (repeatable) --cases FILE one sample per line; "+ text" must match, "- text" must not match, "# ..." is a comment, other lines are shown without an expectation --mode MODE search (default), match, fullmatch, or findall --flags LETTERS any of i m s x a (IGNORECASE, MULTILINE, DOTALL, VERBOSE, ASCII) --probe time the pattern on generated worst-case inputs (subprocess, 2 s limit) --no-explain skip the token-by-token explanation --json print the result as JSON --fail-on LEVEL exit 1 if any risk is at least LEVEL: high (default), warn, info, never --timeout S stop matching one sample after S seconds and report TIMEOUT (default 2) Risks: HIGH = nested quantifiers that can backtrack exponentially; WARN = overlapping alternatives or adjacent quantifiers, [A-z], unescaped dots, empty matches; INFO = anchoring, $ versus \\Z, Unicode \\d and \\w, greedy .*, flags that look missing. Exit codes: 0 ok, 1 failed expectations or a risk at or above --fail-on, 2 usage or pattern error. Python 3 standard library only (uses the re module's own parser). """ import json import re import signal import string import subprocess import sys try: # Python 3.11+ import re._parser as sre_parse except ImportError: # pragma: no cover - older Python import sre_parse # type: ignore LEVELS = {"info": 0, "warn": 1, "high": 2} FLAG_LETTERS = {"i": re.IGNORECASE, "m": re.MULTILINE, "s": re.DOTALL, "x": re.VERBOSE, "a": re.ASCII} FLAG_NAMES = [(re.IGNORECASE, "IGNORECASE"), (re.MULTILINE, "MULTILINE"), (re.DOTALL, "DOTALL"), (re.VERBOSE, "VERBOSE"), (re.ASCII, "ASCII")] ALPHABET = string.ascii_letters + string.digits + string.punctuation + " \t\n" + "\u00e9\u0663" CATEGORY_RE = {"CATEGORY_DIGIT": r"\d", "CATEGORY_NOT_DIGIT": r"\D", "CATEGORY_SPACE": r"\s", "CATEGORY_NOT_SPACE": r"\S", "CATEGORY_WORD": r"\w", "CATEGORY_NOT_WORD": r"\W"} CATEGORY_TEXT = {"CATEGORY_DIGIT": "a digit", "CATEGORY_NOT_DIGIT": "a non-digit", "CATEGORY_SPACE": "whitespace", "CATEGORY_NOT_SPACE": "a non-whitespace char", "CATEGORY_WORD": "a word char (letter, digit, _)", "CATEGORY_NOT_WORD": "a non-word char"} AT_TEXT = {"AT_BEGINNING": "^ start of string (start of each line with MULTILINE)", "AT_BEGINNING_STRING": "\\A start of string only", "AT_END": "$ end of string or before a final newline (end of each line with MULTILINE)", "AT_END_STRING": "\\Z end of string only", "AT_BOUNDARY": "\\b word boundary", "AT_NON_BOUNDARY": "\\B not a word boundary"} MAXREPEAT = sre_parse.MAXREPEAT class UsageError(Exception): pass class MatchTimeout(Exception): pass def with_timeout(seconds, fn, *args): """Run fn with a wall-clock limit (SIGALRM on Unix; no limit elsewhere).""" if not hasattr(signal, "setitimer") or seconds <= 0: return fn(*args) def handler(_sig, _frame): raise MatchTimeout() old = signal.signal(signal.SIGALRM, handler) signal.setitimer(signal.ITIMER_REAL, seconds) try: return fn(*args) finally: signal.setitimer(signal.ITIMER_REAL, 0) signal.signal(signal.SIGALRM, old) def op_name(op): return str(op) def show_char(code): ch = chr(code) named = {"\n": "\\n", "\t": "\\t", "\r": "\\r", " ": "space"} return named.get(ch, repr(ch) if ch.isprintable() else f"U+{code:04X}") # ---------------------------------------------------------------- char sets def class_accepts(items, flags): """Set of ALPHABET chars accepted by an IN [...] item list.""" negate = bool(items) and op_name(items[0][0]) == "NEGATE" acc = set() for op, av in items: name = op_name(op) if name == "LITERAL": acc.add(chr(av)) elif name == "RANGE": acc |= {c for c in ALPHABET if av[0] <= ord(c) <= av[1]} elif name == "CATEGORY": rx = re.compile(CATEGORY_RE.get(op_name(av), "."), flags & re.ASCII) acc |= {c for c in ALPHABET if rx.match(c)} if flags & re.IGNORECASE: acc |= {c.swapcase() for c in acc} return set(ALPHABET) - acc if negate else acc def first_info(seq, flags): """Return (first_chars, can_be_empty) for a token sequence.""" first, empty = set(), True for op, av in seq: f, e = token_first(op, av, flags) first |= f if not e: empty = False break return first, empty def token_first(op, av, flags): name = op_name(op) if name == "LITERAL": c = chr(av) return ({c, c.swapcase()} if flags & re.IGNORECASE else {c}), False if name == "NOT_LITERAL": return set(ALPHABET) - {chr(av)}, False if name == "IN": return class_accepts(av, flags), False if name == "ANY": return (set(ALPHABET) if flags & re.DOTALL else set(ALPHABET) - {"\n"}), False if name in ("MAX_REPEAT", "MIN_REPEAT", "POSSESSIVE_REPEAT"): lo, _hi, sub = av f, e = first_info(sub, flags) return f, e or lo == 0 if name == "SUBPATTERN": return first_info(av[3], flags_for(av, flags)) if name == "ATOMIC_GROUP": return first_info(av, flags) if name == "BRANCH": first, empty = set(), False for b in av[1]: f, e = first_info(b, flags) first |= f empty = empty or e return first, empty if name == "GROUPREF_EXISTS": _g, yes, no = av f1, e1 = first_info(yes, flags) f2, e2 = first_info(no or [], flags) return f1 | f2, e1 or e2 if name == "GROUPREF": return set(ALPHABET), True return set(), True # AT, ASSERT, ASSERT_NOT: zero width def flags_for(av, flags): add, delete = av[1], av[2] return (flags | add) & ~delete # ---------------------------------------------------------------- explanation def quant_text(lo, hi, name): mode = {"MAX_REPEAT": "greedy", "MIN_REPEAT": "lazy", "POSSESSIVE_REPEAT": "possessive"}[name] if hi == MAXREPEAT: base = {0: "zero or more", 1: "one or more"}.get(lo, f"{lo} or more") elif lo == hi: return f"exactly {lo}" elif (lo, hi) == (0, 1): base = "optional (zero or one)" else: base = f"{lo} to {hi}" return f"{base} ({mode})" def class_text(items): negate = bool(items) and op_name(items[0][0]) == "NEGATE" parts = [] for op, av in items: name = op_name(op) if name == "LITERAL": parts.append(show_char(av)) elif name == "RANGE": parts.append(f"{show_char(av[0])}-{show_char(av[1])}") elif name == "CATEGORY": parts.append(CATEGORY_TEXT.get(op_name(av), op_name(av))) return ("any char except " if negate else "one of: ") + ", ".join(parts) def explain(seq, group_names, flags, depth=0, out=None): out = [] if out is None else out pad = " " * depth i = 0 while i < len(seq): op, av = seq[i] name = op_name(op) if name == "LITERAL": # merge runs of literals j = i while j < len(seq) and op_name(seq[j][0]) == "LITERAL": j += 1 text = "".join(chr(a) for _, a in seq[i:j]) out.append(f"{pad}the text {text!r}" if j - i > 1 else f"{pad}the char {show_char(av)}") i = j continue if name == "NOT_LITERAL": out.append(f"{pad}any char except {show_char(av)}") elif name == "IN": out.append(f"{pad}{class_text(av)}") elif name == "ANY": out.append(f"{pad}any char" + ("" if flags & re.DOTALL else " except newline")) elif name == "AT": out.append(f"{pad}{AT_TEXT.get(op_name(av), op_name(av))}") elif name in ("MAX_REPEAT", "MIN_REPEAT", "POSSESSIVE_REPEAT"): lo, hi, sub = av out.append(f"{pad}{quant_text(lo, hi, name)} of:") explain(sub, group_names, flags, depth + 1, out) elif name == "SUBPATTERN": g = av[0] if g is None: out.append(f"{pad}non-capturing group:") else: label = f' "{group_names[g]}"' if g in group_names else "" out.append(f"{pad}capturing group {g}{label}:") explain(av[3], group_names, flags_for(av, flags), depth + 1, out) elif name == "ATOMIC_GROUP": out.append(f"{pad}atomic group (no backtracking into it):") explain(av, group_names, flags, depth + 1, out) elif name == "BRANCH": out.append(f"{pad}either:") for k, b in enumerate(av[1]): out.append(f"{pad} option {k + 1}:") explain(b, group_names, flags, depth + 2, out) elif name in ("ASSERT", "ASSERT_NOT"): direction, sub = av kind = ("lookbehind" if direction < 0 else "lookahead") neg = "negative " if name == "ASSERT_NOT" else "" out.append(f"{pad}{neg}{kind} (checks without consuming):") explain(sub, group_names, flags, depth + 1, out) elif name == "GROUPREF": out.append(f"{pad}the same text as group {av} matched") elif name == "GROUPREF_EXISTS": out.append(f"{pad}if group {av[0]} matched:") explain(av[1], group_names, flags, depth + 1, out) if av[2]: out.append(f"{pad}otherwise:") explain(av[2], group_names, flags, depth + 1, out) else: out.append(f"{pad}{name}") i += 1 return out # ---------------------------------------------------------------- risk analysis def walk_repeats(seq, flags, path=""): """Yield (op, av, flags) for every repeat node, outside lookarounds too.""" for op, av in seq: name = op_name(op) if name in ("MAX_REPEAT", "MIN_REPEAT", "POSSESSIVE_REPEAT"): yield op, av, flags yield from walk_repeats(av[2], flags) elif name == "SUBPATTERN": yield from walk_repeats(av[3], flags_for(av, flags)) elif name == "ATOMIC_GROUP": yield from walk_repeats(av, flags) elif name == "BRANCH": for b in av[1]: yield from walk_repeats(b, flags) elif name in ("ASSERT", "ASSERT_NOT"): yield from walk_repeats(av[1], flags) def flatten(seq): """Expand groups so the body of a repeat can be read as one token list.""" out = [] for op, av in seq: name = op_name(op) if name == "SUBPATTERN" and av[1] == 0 and av[2] == 0: out.extend(flatten(av[3])) else: out.append((op, av)) return out def token_width(tok, parsed): return sre_parse.SubPattern(parsed.state, [tok]).getwidth() def all_chars(seq, flags): """Every ALPHABET char a token sequence can consume anywhere (not only first).""" acc = set() for op, av in seq: name = op_name(op) if name in ("LITERAL", "NOT_LITERAL", "IN", "ANY"): acc |= token_first(op, av, flags)[0] elif name in ("MAX_REPEAT", "MIN_REPEAT", "POSSESSIVE_REPEAT"): acc |= all_chars(av[2], flags) elif name == "SUBPATTERN": acc |= all_chars(av[3], flags_for(av, flags)) elif name == "ATOMIC_GROUP": acc |= all_chars(av, flags) elif name == "BRANCH": for b in av[1]: acc |= all_chars(b, flags) elif name == "GROUPREF_EXISTS": acc |= all_chars(av[1], flags) | all_chars(av[2] or [], flags) elif name == "GROUPREF": acc |= set(ALPHABET) return acc def has_unbounded(tok): op, av = tok name = op_name(op) if name in ("MAX_REPEAT", "MIN_REPEAT"): return av[1] == MAXREPEAT or has_unbounded_seq(av[2]) if name == "SUBPATTERN": return has_unbounded_seq(av[3]) if name == "BRANCH": return any(has_unbounded_seq(b) for b in av[1]) return False def has_unbounded_seq(seq): return any(has_unbounded(t) for t in seq) def sample_char(chars): for pool in (string.ascii_lowercase, string.digits, string.ascii_uppercase, ALPHABET): for c in pool: if c in chars: return c return None def analyze(parsed, flags, pattern, modes_seen): risks = [] def add(level, rule, msg, probe=None): risks.append({"level": level, "rule": rule, "message": msg, **({"probe": probe} if probe else {})}) # nested quantifiers: a repeat whose body can match the same text in more than one way for op, av, fl in walk_repeats(list(parsed), flags): if op_name(op) == "POSSESSIVE_REPEAT" or av[1] <= 1: continue body = flatten(av[2]) widths = [token_width(t, parsed) for t in body] if sum(w[0] for w in widths) == sum(w[1] for w in widths): continue # fixed-width body such as (\d{4}-)+ cannot be split two ways variable = [t for t, w in zip(body, widths) if w[0] != w[1]] fixed = [t for t, w in zip(body, widths) if w[0] == w[1] and w[0] > 0] var_chars = set().union(*(all_chars([t], fl) for t in variable)) has_inner_loop = any(has_unbounded(t) for t in variable) overlap = set().union(*(all_chars([t], fl) for t in fixed)) & var_chars if fixed else set() sample = sample_char(overlap) or sample_char(var_chars) if not fixed and has_inner_loop: add("high", "nested-quantifier", "a quantifier sits inside another quantifier and nothing else in the outer repeat is " "required (the classic (a+)+ or (\\w+\\s?)+ shape), so one string can be split in " "exponentially many ways; rewrite as a single repeat, make the separator required, or use an " "atomic group or possessive quantifier (Python 3.11+)", probe=sample) elif any(not (all_chars([t], fl) & var_chars) for t in fixed): if has_inner_loop: add("info", "nested-separated", "nested quantifier with a required separator that the inner repeat cannot match " "(like ([\\w-]+\\.)+); this shape is safe") elif all(all_chars([t], fl) <= var_chars for t in fixed): add("high", "ambiguous-repeat", "the body of a repeat has variable length and every required part can also be eaten by " "the variable part (for example (x+x+)+, (aa|a)+, or (\\w+|\\d+)+), so the iterations " "can be split in exponentially many ways; merge the parts into one class or make each " "iteration start with a char the rest cannot match", probe=sample) else: add("warn", "nested-overlap", "a repeat with a variable-length body whose separator is also accepted by the inner part " "(for example ([\\w.]+\\.)+): long non-matching input can slow down sharply; make the inner " "class exclude the separator", probe=sample) # adjacent overlapping unbounded repeats in any sequence seen_pairs = set() def single_char_repeat(op, av): return (op_name(op) in ("MAX_REPEAT", "MIN_REPEAT") and av[1] == MAXREPEAT and len(av[2]) == 1 and op_name(av[2][0][0]) in ("LITERAL", "NOT_LITERAL", "IN", "ANY")) def scan_adjacent(seq, fl): toks = [t for t in flatten(seq) if op_name(t[0]) not in ("AT", "ASSERT", "ASSERT_NOT")] for (a_op, a_av), (b_op, b_av) in zip(toks, toks[1:]): if single_char_repeat(a_op, a_av) and single_char_repeat(b_op, b_av): if (id(a_av), id(b_av)) in seen_pairs: continue seen_pairs.add((id(a_av), id(b_av))) fa, _ = first_info(a_av[2], fl) fb, _ = first_info(b_av[2], fl) if fa & fb: add("warn", "adjacent-overlap", "two unbounded quantifiers in a row accept the same chars (for example \\d+\\d+ or " ".*.*), so the split between them is ambiguous and failures backtrack polynomially; " "keep one of them") for op, av in seq: name = op_name(op) if name == "SUBPATTERN": scan_adjacent(av[3], flags_for(av, fl)) elif name in ("MAX_REPEAT", "MIN_REPEAT", "POSSESSIVE_REPEAT"): scan_adjacent(av[2], fl) elif name == "BRANCH": for b in av[1]: scan_adjacent(b, fl) scan_adjacent(list(parsed), flags) # text-level checks if re.search(r"\[[^\]]*A-z", pattern): add("warn", "range-A-z", "[A-z] also matches [ \\ ] ^ _ and the backtick; use [A-Za-z]") seq = list(parsed) flat = flatten(seq) for k, (op, _av) in enumerate(flat): if op_name(op) != "ANY": continue before = flat[k - 1] if k else None after = [chr(a) for o, a in flat[k + 1:k + 4] if op_name(o) == "LITERAL"] lit_before = before is not None and op_name(before[0]) == "LITERAL" and chr(before[1]).isalnum() word_after = len(after) >= 2 and all(c.isalpha() for c in after[:2]) if (lit_before and after and after[0].isalnum()) or word_after: add("warn", "unescaped-dot", f"'.' before {''.join(after) or 'the next token'!r} matches any char (so a '-' or 'x' passes there too); " "write \\. for a literal dot") break try: if re.compile(pattern, flags).fullmatch(""): add("warn", "matches-empty", "the pattern matches the empty string, so findall/finditer return empty matches and " "validation accepts blank input; require at least one char where it matters") except re.error: pass count_dotstar = len(re.findall(r"(?<!\\)\.\*", pattern)) if count_dotstar >= 3: add("info", "many-dotstar", f"{count_dotstar} '.*' in one pattern: failures are slow on long lines; " "use negated classes like [^,]* to describe each field") elif count_dotstar and not re.search(r"\.\*\?", pattern): add("info", "greedy-dotstar", "greedy .* grabs as much as possible and then backs off; use .*? or a " "negated class such as [^\"]* when you mean 'up to the next ...'") first_ops = [op_name(op) for op, _ in seq[:1]] last = seq[-1] if seq else None anchored_start = first_ops == ["AT"] anchored_end = last is not None and op_name(last[0]) == "AT" and op_name(last[1]) in ("AT_END", "AT_END_STRING") if "fullmatch" not in modes_seen and not (anchored_start and anchored_end) and modes_seen != {"findall"}: add("info", "not-anchored", "for validation, anchor both ends (\\A ... \\Z) or use re.fullmatch; " "otherwise 'abc123xyz' passes a search for \\d+") if anchored_end and op_name(last[1]) == "AT_END" and not flags & re.MULTILINE and "fullmatch" not in modes_seen: add("info", "dollar-newline", "$ also matches before a trailing newline ('abc\\n' passes ^abc$); " "use \\Z or fullmatch for strict validation") if re.search(r"(?<!\\)\\[dw]", pattern) and not flags & re.ASCII: add("info", "unicode-classes", "\\d and \\w match Unicode digits and letters too (for example " "Arabic-Indic digits); add re.ASCII or use [0-9] / [A-Za-z0-9_] for strict input formats") return risks # ---------------------------------------------------------------- probe PROBE_CODE = r""" import re, sys, time pat, flags, s = sys.argv[1], int(sys.argv[2]), sys.argv[3] rx = re.compile(pat, flags) t = time.perf_counter(); rx.search(s); print(f"{time.perf_counter() - t:.6f}") """ def run_probe(pattern, flags, ch, compiled): tail = next((t for t in "!#\x00~ \n" if not compiled.search(ch * 3 + t)), "!") results, verdict = [], "no slowdown seen" for n in (12, 16, 20, 24, 28): s = ch * n + tail try: r = subprocess.run([sys.executable, "-c", PROBE_CODE, pattern, str(flags), s], capture_output=True, text=True, timeout=2) secs = float(r.stdout.strip() or "nan") except subprocess.TimeoutExpired: results.append({"n": n, "seconds": None}) verdict = f"TIMEOUT at length {n + 1}: exponential backtracking confirmed" break results.append({"n": n, "seconds": round(secs, 4)}) prev = results[-2]["seconds"] if len(results) >= 2 else None if prev and secs > 0.01 and secs / prev > 3: verdict = (f"time grows about {secs / prev:.0f}x per +4 chars: exponential backtracking confirmed") elif secs > 0.5: verdict = f"slow ({secs:.2f} s at length {n + 1}) but not clearly exponential" return {"input": f"{ch!r} * n + {tail!r}", "timings": results, "verdict": verdict} # ---------------------------------------------------------------- matching def read_cases(path): text = sys.stdin.read() if path == "-" else open(path, encoding="utf-8").read() cases = [] for raw in text.splitlines(): if not raw.strip() or raw.lstrip().startswith("#"): continue if raw.startswith("+ ") or raw.startswith("- "): cases.append({"text": raw[2:], "expect": raw[0] == "+"}) else: cases.append({"text": raw, "expect": None}) return cases def match_case(rx, text, mode): fn = {"search": rx.search, "match": rx.match, "fullmatch": rx.fullmatch}.get(mode) names = {v: k for k, v in rx.groupindex.items()} def groups(m): out = [] for g in range(1, rx.groups + 1): out.append({"group": g, "name": names.get(g), "value": m.group(g), "span": list(m.span(g)) if m.group(g) is not None else None}) return out if mode == "findall": ms = list(rx.finditer(text)) return {"matched": bool(ms), "matches": [{"text": m.group(0), "span": list(m.span()), "groups": groups(m)} for m in ms]} m = fn(text) if not m: return {"matched": False, "matches": []} return {"matched": True, "matches": [{"text": m.group(0), "span": list(m.span()), "groups": groups(m)}]} def describe_flags(flags): names = [n for f, n in FLAG_NAMES if flags & f] return ", ".join(names) if names else "none" # ---------------------------------------------------------------- main def parse_args(argv): opts = {"texts": [], "cases": None, "mode": "search", "flags": 0, "probe": False, "explain": True, "json": False, "fail_on": "high", "pattern": None, "timeout": 2.0} i = 0 while i < len(argv): a = argv[i] def val(): if i + 1 >= len(argv): raise UsageError(f"{a} needs a value") return argv[i + 1] if a in ("-t", "--text"): opts["texts"].append(val()); i += 2 elif a == "--cases": opts["cases"] = val(); i += 2 elif a == "--mode": opts["mode"] = val(); i += 2 if opts["mode"] not in ("search", "match", "fullmatch", "findall"): raise UsageError("--mode must be search, match, fullmatch, or findall") elif a == "--flags": letters = val(); i += 2 bad = [c for c in letters.lower() if c not in FLAG_LETTERS] if bad: raise UsageError(f"unknown flag letter(s) {''.join(bad)!r}; use any of i m s x a") for c in letters.lower(): opts["flags"] |= FLAG_LETTERS[c] elif a == "--fail-on": opts["fail_on"] = val().lower(); i += 2 if opts["fail_on"] not in ("high", "warn", "info", "never"): raise UsageError("--fail-on must be high, warn, info, or never") elif a == "--timeout": try: opts["timeout"] = float(val()) except ValueError: raise UsageError("--timeout needs a number of seconds, for example --timeout 2") i += 2 elif a == "--probe": opts["probe"] = True; i += 1 elif a == "--no-explain": opts["explain"] = False; i += 1 elif a == "--json": opts["json"] = True; i += 1 elif a in ("-h", "--help"): print(__doc__); sys.exit(0) elif a.startswith("--") and opts["pattern"] is not None: raise UsageError(f"unknown option {a}") elif opts["pattern"] is None: opts["pattern"] = a; i += 1 else: raise UsageError(f"unexpected argument {a!r} (quote the pattern; use -t for sample inputs)") if opts["pattern"] is None: raise UsageError("give a PATTERN (quote it in single quotes in the shell)") return opts def main(argv): try: opts = parse_args(argv) cases = [{"text": t, "expect": None} for t in opts["texts"]] if opts["cases"]: try: cases += read_cases(opts["cases"]) except OSError as e: raise UsageError(f"cannot read cases file: {e.strerror}: {opts['cases']}") except UsageError as e: print(f"error: {e}", file=sys.stderr) return 2 pattern, flags, mode = opts["pattern"], opts["flags"], opts["mode"] try: rx = re.compile(pattern, flags) parsed = sre_parse.parse(pattern, flags) except re.error as e: pos = f" at position {e.pos}" if e.pos is not None else "" print(f"error: pattern does not compile: {e.msg}{pos}", file=sys.stderr) if e.pos is not None: print(f" {pattern}\n {' ' * e.pos}^", file=sys.stderr) return 2 flags = int(rx.flags) & ~int(re.UNICODE) names = {v: k for k, v in rx.groupindex.items()} risks = analyze(parsed, flags, pattern, {mode}) if opts["probe"]: for r in risks: if r.get("probe"): r["probe_result"] = run_probe(pattern, flags, r["probe"], rx) results = [] for c in cases: try: res = with_timeout(opts["timeout"], match_case, rx, c["text"], mode) status = None if c["expect"] is None else ("PASS" if res["matched"] == c["expect"] else "FAIL") except MatchTimeout: res, status = {"matched": None, "matches": []}, "TIMEOUT" results.append({**c, **res, "status": status}) fails = sum(1 for r in results if r["status"] in ("FAIL", "TIMEOUT")) counts = {lvl: sum(1 for r in risks if r["level"] == lvl) for lvl in ("high", "warn", "info")} report = {"pattern": pattern, "flags": describe_flags(flags), "mode": mode, "groups": rx.groups, "group_names": {str(k): v for k, v in names.items()}, "explanation": explain(list(parsed), names, flags) if opts["explain"] else [], "cases": results, "failed_expectations": fails, "risks": risks, "risk_counts": counts} threshold = LEVELS.get(opts["fail_on"], 99) risky = any(LEVELS[r["level"]] >= threshold for r in risks) if opts["json"]: print(json.dumps(report, indent=2, ensure_ascii=False)) return 1 if fails or risky else 0 print(f"Pattern: {pattern}") named = ", ".join(f'{g} "{n}"' for g, n in sorted(names.items())) print(f"Flags: {describe_flags(flags)} | mode: {mode} | groups: {rx.groups}" + (f" ({named})" if named else "")) if opts["explain"]: print("\nExplanation:") for line in report["explanation"]: print(" " + line) if results: print(f"\nSamples ({mode}):") for r in results: mark = {True: "+", False: "-", None: " "}[r["expect"]] tag = r["status"] or ("match" if r["matched"] else "no match") line = f" {tag:<8} {mark} {r['text']!r}" if r["matched"]: parts = [] for m in r["matches"][:5]: g = " ".join(f"{x['name'] or x['group']}={x['value']!r}" for x in m["groups"]) parts.append(f"[{m['span'][0]}:{m['span'][1]}] {m['text']!r}" + (f" {g}" if g else "")) more = f" (+{len(r['matches']) - 5} more)" if len(r["matches"]) > 5 else "" line += " -> " + "; ".join(parts) + more elif r["status"] == "TIMEOUT": line += f" -> still running after {opts['timeout']:g} s (catastrophic backtracking)" elif r["status"] == "FAIL": line += " -> expected a match" if r["status"] == "FAIL" and r["matched"]: line += " (expected no match)" print(line) print("\nRisks and suggestions:") if not risks: print(" none found") for r in risks: print(f" {r['level'].upper():<4} {r['rule']:<23} {r['message']}") pr = r.get("probe_result") if pr: times = ", ".join(f"n={t['n']}: {'timeout' if t['seconds'] is None else str(t['seconds']) + ' s'}" for t in pr["timings"]) print(f" probe {pr['input']}: {times} -> {pr['verdict']}") summary = f"\nSummary: HIGH {counts['high']} WARN {counts['warn']} INFO {counts['info']}" if any(r["expect"] is not None for r in results): n = sum(1 for r in results if r["expect"] is not None) summary += f" | expectations {n - fails}/{n} pass" timeouts = sum(1 for r in results if r["status"] == "TIMEOUT") if timeouts: summary += f" | {timeouts} sample(s) timed out" print(summary) return 1 if fails or risky else 0 if __name__ == "__main__": sys.exit(main(sys.argv[1:]))
Turns a rough podcast transcript, timestamps, and guest notes into publish-ready show notes: an episode title, an SEO summary, key takeaways, chapter markers for YouTube, Spotify, and Podcasting 2.0, timed pull quotes, a guest bio, links, and social posts. Use when a user shares a transcript and asks "write my show notes", "make chapters for this episode", or "pull quotes from this interview". Includes a tested stdlib Python chapter finder.
---
name: podcast-show-notes-structurer
description: Turns a rough podcast transcript, timestamps, and guest notes into publish-ready show notes: an episode title, an SEO summary, key takeaways, chapter markers for YouTube, Spotify, and Podcasting 2.0, timed pull quotes, a guest bio, links, and social posts. Use when a user shares a transcript and asks "write my show notes", "make chapters for this episode", or "pull quotes from this interview". Includes a tested stdlib Python chapter finder.
---
# Podcast Show Notes Structurer
You help podcasters turn a messy transcript into show notes that listeners actually read and search engines can find. You find the real structure of the conversation, name chapters the way a listener would describe them, choose quotes that work out of context, and write a summary that tells someone in two lines why this episode is worth their time. You never invent facts, links, or quotes that are not in the material.
## Files in this skill
- `scripts/chapterize.py` - reads timestamped, WebVTT, headed, or plain transcripts, proposes chapter starts from headings, pause markers, cue phrases ("let's talk about", "after the break"), and vocabulary shifts, drafts keyword titles, lists ad and music markers, and ranks quote candidates (Python 3 standard library only)
- `references/show-notes-structure.md` - what goes in each section, length limits, and how to write titles, summaries, and takeaways
- `references/chapters-and-platforms.md` - chapter rules for YouTube, Spotify, Apple Podcasts, and Podcasting 2.0, plus quote and timestamp ethics
- `templates/show-notes.md` - the show notes document to return
- `examples/example-rooftop-beekeeping-episode.md` - a worked example from a 25-minute interview
## Workflow
### 1. Collect the material
Ask, or assume and say so:
- The transcript (any format) and whether its times match the final edit. If an intro or ad was added after transcription, get the offset in seconds.
- Show name, host and guest names with correct spelling, the guest's role, and links they want included (only links the user gives you).
- Where the notes will be published (podcast host, YouTube, website, newsletter) and any house style: tone, length, emoji, sponsor wording.
- The audience and 2 or 3 search phrases a listener might type to find this episode.
### 2. Run the chapter finder
```bash
python3 scripts/chapterize.py transcript.txt
python3 scripts/chapterize.py transcript.txt --min-chapter 3 --offset 45 --duration 52:30
python3 scripts/chapterize.py episode.vtt --format youtube
python3 scripts/chapterize.py - --format json < transcript.txt
```
The script's chapter starts and quotes are candidates, not answers. Read the transcript around every proposed start and move it to the sentence where the new topic really begins. Rewrite every keyword title (for example "Roof / Hive / Water / Wind") into a short listener-facing title ("Keeping bees on a windy roof"). If there are no timestamps, times are estimates (marked `~`); tell the user to replace them with times from the audio editor. If you cannot run the script, find chapters by hand the same way: headings, breaks, transition phrases, then topic changes.
### 3. Write the notes
1. Title: 40 to 60 characters, the main promise of the episode, guest name if it helps search.
2. Summary: 2 sentences for the app preview (the first 150 characters must work alone), then one short paragraph.
3. Key takeaways: 3 to 5 bullets, each a concrete idea from the episode, not a topic name.
4. Chapters: 4 to 10 for a 20 to 60 minute episode, first at 00:00, titles under 40 characters, in the order they happen.
5. Quotes: 3 to 5 lines that make sense without context, with speaker and time. Remove filler words only; never change meaning. Mark edited quotes so the user can check them against the audio.
6. Guest bio (2 or 3 sentences), links and resources mentioned (only real ones from the material or the user), credits, sponsor disclosure if there was an ad.
7. Social posts: one per platform the user names, each with one quote or takeaway.
### 4. Deliver
Fill in `templates/show-notes.md` and follow `references/show-notes-structure.md` and `references/chapters-and-platforms.md`. Follow the style of `examples/example-rooftop-beekeeping-episode.md`. Add a short "check before publishing" list: estimated times, edited quotes, spellings, and any facts or numbers the guest stated that should be verified.
## Rules
- Only use facts, names, numbers, and links from the transcript or the user; flag anything unclear as [check].
- Quotes must be the speaker's own words; removing "um" and "you know" is fine, rewording is not.
- Keep chapter titles descriptive and honest; no clickbait the episode does not deliver.
- Respect guests: do not quote private remarks, off-the-record moments, or anything the user marks as cut.
- Match the show's voice; default to warm, clear, and specific.
FILE:references/show-notes-structure.md
# Show notes structure
Listeners skim show notes in a podcast app on a phone, search engines read them on the episode page, and the guest shares them. Each section has one job.
| Section | Job | Length |
|---|---|---|
| Episode title | Say the main promise; help search | 40 to 60 characters; most apps cut after about 60 |
| Preview summary | Make someone press play from the app list | First 150 characters must work alone; 2 sentences total |
| Full summary | Explain who the guest is and what you cover | 60 to 120 words |
| Key takeaways | Give skimmers the value | 3 to 5 bullets, one idea each, under 20 words |
| Chapters | Let listeners jump | 4 to 10 for 20 to 60 minutes; titles under 40 characters |
| Pull quotes | Quotable lines for notes and social | 3 to 5, 12 to 30 words each, with speaker and time |
| Guest bio | Credit the guest | 2 or 3 sentences, written in third person |
| Links and resources | Everything mentioned that has a link | Only links the user or guest gave |
| Credits and disclosure | Host, guest, producer, music, sponsor | One line each |
| Call to action | One next step | Subscribe, review, newsletter, or a question for listeners |
## Titles
- Lead with the idea, not the episode number: "Keeping Bees on a Hotel Roof with Daniel Okafor" beats "Ep. 48: Daniel Okafor".
- Use the words a listener would search: "rooftop beekeeping", "city honey".
- Avoid questions the episode does not answer and promises the guest did not make.
## Summaries
- Sentence 1: who and what, in plain words. Sentence 2: the most surprising or useful point.
- Write for people, not keywords; include the 2 or 3 search phrases naturally once each.
- No "In this episode we talk about..." filler; start with the substance.
## Key takeaways
Write each takeaway as a claim or instruction from the episode, not a topic:
- Weak: "Water on roofs."
- Strong: "Give rooftop bees a shallow water tray, or they will find the neighbors' pool."
## Chapter titles
- Describe what the listener gets in that part, in 2 to 6 words: "Why wind is the real enemy".
- Keep them parallel in style (all statements or all questions).
- Name ad breaks only if the platform does not hide them; never call an ad a chapter about content.
## Quotes
- Pick lines that stand alone: a clear claim, a lesson, a number, or a vivid image.
- Avoid greetings, setup questions, and lines that need the previous sentence.
- Removing fillers (um, uh, you know) is fine; mark the quote as lightly edited. Changing words is not.
## Accessibility and trust
- Link a full transcript when one exists; it helps deaf and hard-of-hearing listeners and search.
- Disclose sponsors and affiliate links clearly.
- Put a [check] tag on any number, name, or claim that should be verified before publishing.
FILE:references/chapters-and-platforms.md
# Chapters and platform rules
Rules change; when in doubt, check the platform's current help page. These are the stable basics.
## YouTube (video podcasts and audio uploads)
- Chapters come from timestamps in the description, one per line: `00:00 Intro`.
- The first timestamp must be 00:00.
- At least 3 timestamps, and each chapter at least 10 seconds long.
- Use `MM:SS` for videos under an hour and `H:MM:SS` once any chapter is past one hour.
- `chapterize.py --format youtube` prints lines in this format and warns when a rule is broken.
## Spotify
- Spotify reads chapter timestamps from the episode description in the same `MM:SS Title` style for episodes on its platform; follow the same rules as YouTube (start at 00:00, a few minutes apart).
## Apple Podcasts and most other apps
- Chapters are embedded in the audio file: ID3 CHAP frames in MP3 or a chapter track in M4A/AAC. Add them in the audio editor or a tagging tool before uploading.
- Apple may show automatically generated chapters when an episode has none; providing your own keeps control of the titles.
## Podcasting 2.0 (RSS)
- The `<podcast:chapters>` tag points to a JSON chapters file, supported by many modern apps and hosts:
```json
{
"version": "1.2.0",
"chapters": [
{ "startTime": 0, "title": "Intro" },
{ "startTime": 280, "title": "Keeping bees on a windy roof" }
]
}
```
- `startTime` is in seconds; `chapterize.py --format json` gives `start` in seconds for each chapter.
## Timestamps
- Transcript times often differ from the final file: intros, ads, and edits shift them. Use `--offset` for a fixed shift and recheck a few chapters in the audio.
- Without timestamps, `chapterize.py` estimates times from word count (default 150 words per minute) and marks them with `~`. Never publish estimated times without checking.
- Dynamic ad insertion moves times per listener; put chapters in the file or description only when ads are baked in or the host adjusts chapters automatically.
## How the script finds chapter starts
1. Headings (`# Title`, `Chapter 3: ...`, `Segment: ...`) always start a chapter.
2. Pause markers (`[music]`, `[ad break]`, `[pause]`, `***`, `---`) mark the next spoken block as a candidate.
3. Section cue phrases score high ("let's talk about", "moving on", "switching gears", "welcome back", "before we wrap up"); single-question cues ("next question") score lower so a Q&A segment stays one chapter.
4. Vocabulary shift: how different the words are in the 3 blocks before and after a point.
5. Starts closer than `--min-chapter` minutes are dropped (headings excepted), and nothing starts in the last minute.
## Quote and transcript ethics
- Keep quotes faithful; mark filler removal as an edit.
- Ask before quoting a guest's personal story on social media if it was sensitive.
- Do not publish remarks the user marked as off the record or cut from the episode.
FILE:templates/show-notes.md
# episode_title
**show_name** | Episode episode_number | duration | Guest: guest_name
preview_summary_two_sentences
full_summary_paragraph
## Key takeaways
-
-
-
## Chapters
```
00:00 chapter_1
time chapter_2
```
## Quotes
> "quote" - speaker (time)
## About guest_name
guest_bio
## Links and resources mentioned
- resource - url_or_check
## Credits
Host: host | Guest: guest_name | Produced by: producer | Music: music_credit
sponsor_disclosure
call_to_action
---
## Social posts
- **platform:** post_with_quote_or_takeaway
## Check before publishing
- [ ] Chapter times checked against the final audio (offset: offset)
- [ ] Edited quotes checked against the audio
- [ ] Names and spellings confirmed
- [ ] Facts and numbers marked [check] verified
- [ ] Links tested
FILE:examples/example-rooftop-beekeeping-episode.md
# Example: show notes for a 25-minute interview
## User request
"Here's the transcript of episode 48 of Small Plot Radio with Daniel Okafor, who keeps bees on a hotel roof. We added a 45-second intro in the edit, so the final file runs 26:02. Can you write the show notes with chapters for YouTube and Spotify, a few quotes, and an Instagram caption? The hotel tours and Daniel's course have links on their sites; I'll add them."
## Chapter finder run
```
$ python3 scripts/chapterize.py rooftop-bees.txt
Transcript: 1140 words, about 25:17 long, timestamps found, speakers: Daniel (20 turns), Maya (19 turns)
Chapter candidates (6, min 4 min apart):
1. 00:00 4.7 min [detected] Hotel Manager / Hive / Twelve / Bees
2. 04:40 4.5 min [detected] Roof / Hive / Water / Wind
3. 09:12 5.3 min [detected] Honey Tour / Spring Honey / Late Summer / Guests
4. 14:30 5.0 min [detected] City / Bees / Honey / Roof
5. 19:30 4.7 min [detected] Hive / Balcony / Bees / Priya
6. 24:10 1.1 min [detected] Keeping / Colony / Association / Find
Markers (breaks, music, ads): 13:30 [ad break]
Quote candidates (6):
00:36 Daniel: "I built my first hive from a kit that winter and I made every beginner mistake you can imagine."
01:24 Daniel: "The biggest lesson I learned that first year was that good beekeeping is mostly patience."
05:33 Daniel: "If you don't give them water, they will find the hotel pool or the neighbors' air conditioning drip, and that is how complaints start."
...
$ python3 scripts/chapterize.py rooftop-bees.txt --offset 45 --format youtube
00:00 Hotel Manager / Hive / Twelve / Bees
05:25 Roof / Hive / Water / Wind
09:57 Honey Tour / Spring Honey / Late Summer / Guests
15:15 City / Bees / Honey / Roof
20:15 Hive / Balcony / Bees / Priya
24:55 Keeping / Colony / Association / Find
# note: YouTube needs the first chapter at 00:00; the first one was moved to 00:00
```
Review of the candidates: all six starts land on real transitions ("Let's talk about the roof itself", "Switching gears to the hotel manager", "Welcome back", "Let's move on to listener questions", "Before we wrap up"). The 13:30 ad break stays inside chapter 3 because the ad is baked in. Chapter 1 is long, so it was split by hand at 01:20 ("What was the biggest mistake?"), which becomes 02:05 after the offset. Every keyword title was rewritten.
## Show notes (filled template)
# Keeping Bees on a Hotel Roof with Daniel Okafor
**Small Plot Radio** | Episode 48 | 26:02 | Guest: Daniel Okafor
Daniel Okafor keeps twelve beehives on a downtown hotel roof. He explains how to beat wind, heat, and thirst on a roof, and why planting flowers can help pollinators more than adding hives.
Nine years ago a swarm landed on Daniel's balcony; today he runs twelve hives on a hotel roof, supplies the hotel kitchen with honey, and leads sold-out rooftop honey tours. Maya and Daniel talk about beginner mistakes, windbreaks, water trays, and roof loads, how he won over a hotel manager, whether city honey is safe to eat, and why too many city hives can crowd out wild bees. They finish with listener questions on balcony beekeeping, starter equipment, and honey yields.
## Key takeaways
- Inspect hives every 7 to 10 days in spring, with a plan, and leave them alone in between.
- On a roof, wind is the main enemy: a 1.5 m slatted windbreak on the exposed sides makes the difference.
- Give rooftop bees a shallow water tray with pebbles, or they will find the neighbors' pool.
- Have the roof load checked first: a full hive can weigh 80 to 100 kg.
- Planting flowers helps pollinators more than adding another hive to a crowded city.
## Chapters
```
00:00 How a swarm made Daniel a beekeeper
02:05 The beginner mistake: opening the hive too often
05:25 Keeping bees on a windy roof
09:57 Winning over the hotel (and its kitchen)
15:15 Is city honey safe, and are there too many hives?
20:15 Listener questions: balconies, gear, and honey yields
24:55 The one thing every new beekeeper should know
```
## Quotes
> "The biggest lesson I learned that first year was that good beekeeping is mostly patience." - Daniel Okafor (02:09)
> "If you don't give them water, they will find the hotel pool or the neighbors' air conditioning drip, and that is how complaints start." - Daniel Okafor (06:18)
> "If you want to help pollinators, planting flowers matters more than adding another hive." - Daniel Okafor (16:53)
## About Daniel Okafor
Daniel Okafor is a city beekeeper who has kept bees for nine years and now runs twelve hives on the roof of a downtown hotel. He leads the hotel's monthly rooftop honey tours and teaches a beginner course with the city beekeeping association every spring.
## Links and resources mentioned
- Hotel rooftop honey tours - [link from user]
- Beginner beekeeping course, city beekeeping association - [link from user]
- Mason bee houses for balconies (mentioned at 20:15) - no link given
## Credits
Host: Maya | Guest: Daniel Okafor | Produced by: [check] | Music: [check]
This episode includes an advertisement at 14:15. [check sponsor name for disclosure]
Got a balcony or roof question for Daniel? Send it in and we'll answer it in a future episode.
---
## Social posts
- **Instagram:** "Planting flowers matters more than adding another hive." Daniel Okafor keeps twelve hives on a hotel roof, and his advice for helping city pollinators might surprise you. New episode of Small Plot Radio, link in bio.
## Check before publishing
- [ ] Chapter times checked against the final audio (offset: +45 s added to every transcript time)
- [ ] Quotes checked against the audio (none were edited)
- [ ] Spelling of "Okafor" confirmed with the guest
- [ ] Facts to verify: "zero stings on guests in nine seasons", yearly lab tests below limits, 15 to 30 kg per hive
- [ ] Links added and tested
FILE:scripts/chapterize.py
#!/usr/bin/env python3
"""Split a podcast transcript into rough chapter candidates and pull quote candidates.
Usage:
python3 chapterize.py TRANSCRIPT [options]
python3 chapterize.py - < transcript.txt (read stdin)
Accepted input (detected automatically, can be mixed):
- timestamp lines: "[00:12:30] Ana: text", "00:12:30 text", "(12:30) text", "12:30 - text"
- WebVTT/SRT cue times: "00:01:02.000 --> 00:01:05.000" (the cue text follows)
- headings: "# Intro", "## Part 2", "Chapter 3: Pricing", "SEGMENT: listener questions"
- pause markers: [pause], [music], [break], [ad break], (long pause), ***, ---
- plain paragraphs separated by blank lines (times are estimated from --wpm)
Options:
--min-chapter MIN shortest chapter in minutes (default 4)
--max-chapters N most chapters to return (default 12)
--wpm N speaking rate used to estimate times without timestamps (default 150)
--offset SEC seconds to add to every time, for an intro added in the edit (default 0)
--quotes N number of quote candidates (default 6)
--format FMT text (default), json, or youtube (chapter lines "00:00 Title")
--duration HH:MM:SS total episode length, to check the last chapter and YouTube rules
Chapter boundaries come from headings first, then pause markers and cue phrases
("let's talk about", "moving on", "after the break"), then vocabulary shifts between
neighbouring blocks. Titles are keyword guesses: always rewrite them for the listener.
Exit codes: 0 ok, 1 input too short for chapters, 2 usage or read error.
Python 3 standard library only.
"""
import json
import re
import sys
from collections import Counter
TS_LINE = re.compile(r"^\s*[\[(]?((?:\d{1,2}:)?\d{1,2}:\d{2})(?:[.,]\d+)?[\])]?\s*(?:[-\u2013\u2014|:]\s*)?(.*)$")
CUE_LINE = re.compile(r"^\s*((?:\d{1,2}:)?\d{1,2}:\d{2})[.,]\d{1,3}\s*-->\s*((?:\d{1,2}:)?\d{1,2}:\d{2})?")
LAST_CUE_END = [0]
HEADING = re.compile(r"^\s*(?:#{1,6}\s+(.+)|(?:chapter|part|segment|section|topic)\s*\d*\s*[:.\-\u2013]\s*(.+))\s*$", re.I)
PAUSE = re.compile(r"^\s*(?:\[(?:pause|long pause|music|break|ad break|ad|sponsor|interlude|silence)[^\]]*\]|"
r"\((?:pause|long pause|music|break|ad break|silence)[^)]*\)|\*{3,}|-{3,}|={3,})\s*$", re.I)
SPEAKER = re.compile(r"^([A-Z][\w.'-]*(?: [A-Z][\w.'-]*){0,2}):\s+(.*)$")
# section-level cue phrases start a new chapter; item-level ones (single questions) only nudge
CUES = ["let's talk about", "lets talk about", "let's move on", "moving on", "switching gears", "change gears",
"after the break", "welcome back", "before we wrap", "to wrap up", "let's get into", "let's dive into",
"let's start with", "rapid fire", "listener questions"]
ITEM_CUES = ["next question", "first question", "last question", "final question", "the next thing",
"another thing i want", "speaking of"]
INTRO = re.compile(r"\b(welcome (to|back)|my guest|today i'm talking|today i am talking|thanks for (having|listening)|"
r"see you (next|in)|where can people find)\b", re.I)
STOP = set("""a about above after again against all also am an and any are aren't as at be because been before being
below between both but by can can't could couldn't did didn't do does doesn't doing don't down during each even ever
every few for from further get gets getting go going gonna got had hadn't has hasn't have haven't having he he'd he'll
he's her here here's hers herself him himself his how how's i i'd i'll i'm i've if in into is isn't it it's its itself
just know kind kinda let let's like lot really right said say says see so some such sort than that that's the their
theirs them themselves then there there's these they they'd they'll they're they've thing things think this those
though through to too um uh under until up us very want was wasn't way we we'd we'll we're we've well were weren't what
what's when when's where where's which while who who's whom why why's will with won't would wouldn't yeah yes you you'd
you'll you're you've your yours yourself yourselves okay ok actually basically maybe mean pretty probably stuff sure
talk talking yeah one two three lot lots much many more most other others same back new good great time people
something anything everything nothing make made making take took come came look looking need needs needed thank
thanks today episode podcast show welcome not first question questions listener listeners she she's she'd she'll now""".split())
QUOTE_WORDS = re.compile(r"\b(never|always|nobody|everyone|the secret|the key|the biggest|the hardest|the best|the worst|"
r"mistake|learned|lesson|truth|matters|changed|surprised|realized|if you|you have to|don't)\b", re.I)
FILLER = re.compile(r"\b(um+|uh+|you know|i mean|sort of|kind of|like,)\s*", re.I)
class UsageError(Exception):
pass
def to_sec(ts):
parts = [int(p) for p in ts.split(":")]
while len(parts) < 3:
parts.insert(0, 0)
h, m, s = parts
return h * 3600 + m * 60 + s
def fmt(sec, force_hours=False):
sec = int(round(sec))
h, rem = divmod(sec, 3600)
m, s = divmod(rem, 60)
return f"{h:d}:{m:02d}:{s:02d}" if h or force_hours else f"{m:02d}:{s:02d}"
def words(text):
return re.findall(r"[A-Za-z][A-Za-z'-]+", text)
def parse(text):
"""Return a list of units: dict(kind=text|heading|pause, text, time or None, speaker)."""
units, pending_time = [], None
lines = text.replace("\r\n", "\n").replace("\r", "\n").lstrip("\ufeff").split("\n")
para = []
def flush():
if para:
joined = " ".join(p["text"] for p in para)
units.append({"kind": "text", "text": joined, "time": para[0]["time"], "speaker": para[0]["speaker"]})
para.clear()
for raw in lines:
line = raw.strip()
if not line or line.upper() == "WEBVTT" or line.isdigit():
if not line:
flush()
continue
m = CUE_LINE.match(line)
if m:
pending_time = to_sec(m.group(1))
if m.group(2):
LAST_CUE_END[0] = max(LAST_CUE_END[0], to_sec(m.group(2)))
continue
h = HEADING.match(line)
if h:
flush()
units.append({"kind": "heading", "text": (h.group(1) or h.group(2)).strip(), "time": pending_time,
"speaker": None})
continue
if PAUSE.match(line):
flush()
units.append({"kind": "pause", "text": line, "time": pending_time, "speaker": None})
continue
t = TS_LINE.match(line)
time = pending_time
if t and t.group(2) is not None and (":" in t.group(1)):
time, line = to_sec(t.group(1)), t.group(2).strip()
flush()
if not line:
pending_time = time
continue
h = HEADING.match(line)
if h:
units.append({"kind": "heading", "text": (h.group(1) or h.group(2)).strip(), "time": time, "speaker": None})
pending_time = None
continue
if PAUSE.match(line):
units.append({"kind": "pause", "text": line, "time": time, "speaker": None})
pending_time = None
continue
sp = SPEAKER.match(line)
speaker = None
if sp and len(sp.group(1)) <= 30:
speaker, line = sp.group(1), sp.group(2)
flush()
if time is not None or pending_time is not None:
flush()
para.append({"text": line, "time": time, "speaker": speaker})
pending_time = None
flush()
return units
def assign_times(units, wpm, offset):
"""Fill missing times by interpolating between known ones or from the word count."""
known = [i for i, u in enumerate(units) if u["time"] is not None]
clock = 0.0
timed = bool(known)
for i, u in enumerate(units):
if u["time"] is not None:
clock = float(u["time"])
u["estimated"] = False
else:
u["time"] = clock
u["estimated"] = True
clock = u["time"] + len(words(u["text"])) / wpm * 60 if u["kind"] == "text" else u["time"]
for u in units:
u["time"] += offset
return timed
EXTRA_STOP = set()
def keywords(text, n=4):
ws = [w.lower().strip("'-") for w in words(text)]
ws = [w for w in ws if len(w) > 2 and w not in STOP and w not in EXTRA_STOP]
counts = Counter(ws)
bigrams = Counter(f"{a} {b}" for a, b in zip(ws, ws[1:]) if a != b)
out = [b for b, c in bigrams.most_common(3) if c >= 2]
for w, _c in counts.most_common(12):
if len(out) >= n:
break
if not any(w in b.split() for b in out):
out.append(w)
return out[:n]
def similarity(a, b):
wa = {w.lower() for w in words(a) if w.lower() not in STOP and len(w) > 2}
wb = {w.lower() for w in words(b) if w.lower() not in STOP and len(w) > 2}
if not wa or not wb:
return 0.0
return len(wa & wb) / len(wa | wb)
def boundaries(units, min_gap, max_chapters, end_time):
"""Score each unit as a chapter start; return chosen indexes (always includes the first text unit)."""
scores = {}
for i, u in enumerate(units):
if u["kind"] == "heading":
scores[i] = 10.0
elif u["kind"] == "pause":
nxt = next((j for j in range(i + 1, len(units)) if units[j]["kind"] == "text"), None)
if nxt is not None:
scores[nxt] = max(scores.get(nxt, 0), 3.0)
elif u["kind"] == "text":
low = u["text"].lower()
cue = 4.0 if any(c in low[:160] for c in CUES) else 1.5 if any(c in low[:160] for c in ITEM_CUES) else 0.0
before = " ".join(x["text"] for x in units[max(0, i - 3):i] if x["kind"] == "text")
after = " ".join(x["text"] for x in units[i:i + 3] if x["kind"] == "text")
shift = 1.0 - similarity(before, after) if before and after else 0.0
score = shift * 2.0 + cue
if score > 1.6:
scores[i] = max(scores.get(i, 0), score)
first = next((i for i, u in enumerate(units) if u["kind"] in ("text", "heading")), 0)
chosen = [first]
for i in sorted(scores, key=lambda k: (-scores[k], k)):
if len(chosen) >= max_chapters:
break
t = units[i]["time"]
forced = units[i]["kind"] == "heading"
strong = forced or scores[i] >= 4.0 # heading or cue phrase
if end_time - t < (60 if strong else min_gap / 2) and not forced:
continue # too close to the end of the episode
if all(abs(t - units[c]["time"]) >= min_gap for c in chosen) or (forced and i not in chosen and
all(abs(t - units[c]["time"]) >= 10 for c in chosen)):
chosen.append(i)
return sorted(set(chosen)), scores
def sentences(text):
return [s.strip() for s in re.split(r"(?<=[.!?])\s+", text) if s.strip()]
def quote_candidates(units, n):
cands = []
for u in units:
if u["kind"] != "text":
continue
for s in sentences(u["text"]):
clean = re.sub(r"\s{2,}", " ", FILLER.sub("", s)).strip()
wc = len(clean.split())
if wc < 8 or wc > 40 or clean.endswith("?"):
continue
score = 0.0
score += 2.0 if 12 <= wc <= 28 else 0.5
score += 1.5 * len(QUOTE_WORDS.findall(clean))
score += 1.0 if re.search(r"\b\d+(?:[.,]\d+)?\s*(%|percent|x|times|years|hours|days)\b", clean, re.I) else 0
score += 0.5 if re.search(r"\b(I|we)\b", clean) else 0
score -= 1.0 if FILLER.search(s) else 0
score -= 3.0 if INTRO.search(clean) else 0
cands.append({"time": u["time"], "speaker": u["speaker"], "text": clean, "score": round(score, 1),
"edited": clean != s})
cands.sort(key=lambda c: (-c["score"], c["time"]))
out, seen = [], set()
for c in cands:
key = c["text"][:40].lower()
if key in seen:
continue
seen.add(key)
out.append(c)
if len(out) >= n:
break
return sorted(out, key=lambda c: c["time"])
def build(units, chosen, end_time):
chapters = []
for k, start in enumerate(chosen):
stop = chosen[k + 1] if k + 1 < len(chosen) else len(units)
seg = units[start:stop]
text = " ".join(u["text"] for u in seg if u["kind"] == "text")
heading = next((u["text"] for u in seg[:1] if u["kind"] == "heading"), None)
t0 = units[start]["time"]
t1 = units[stop]["time"] if stop < len(units) else end_time
speakers = Counter(u["speaker"] for u in seg if u["speaker"])
first = sentences(text)[:1]
chapters.append({
"n": k + 1, "start": round(t0), "start_hms": fmt(t0), "end": round(t1), "minutes": round((t1 - t0) / 60, 1),
"estimated_time": any(u.get("estimated") for u in seg[:1]),
"source": "heading" if heading else "detected",
"title_draft": heading or " / ".join(keywords(text)).title() or "(untitled)",
"keywords": keywords(text, 6), "words": len(words(text)),
"speakers": [s for s, _ in speakers.most_common()],
"opening_line": first[0][:160] if first else "",
})
return chapters
def youtube_lines(chapters):
lines, problems = [], []
if chapters and chapters[0]["start"] != 0:
problems.append("YouTube needs the first chapter at 00:00; the first one was moved to 00:00")
for c in chapters:
start = 0 if c is chapters[0] else c["start"]
lines.append(f"{fmt(start, force_hours=chapters[-1]['start'] >= 3600)} {c['title_draft']}")
if len(chapters) < 3:
problems.append("YouTube shows chapters only with at least 3 of them")
short = [c["n"] for c in chapters if c["minutes"] * 60 < 10]
if short:
problems.append(f"chapters {short} are shorter than 10 seconds (YouTube minimum)")
return lines, problems
def parse_args(argv):
opts = {"file": None, "min_chapter": 4.0, "max_chapters": 12, "wpm": 150.0, "offset": 0.0, "quotes": 6,
"format": "text", "duration": None}
numeric = {"--min-chapter": ("min_chapter", float), "--max-chapters": ("max_chapters", int),
"--wpm": ("wpm", float), "--offset": ("offset", float), "--quotes": ("quotes", int)}
i = 0
while i < len(argv):
a = argv[i]
if a in ("-h", "--help"):
print(__doc__)
sys.exit(0)
if a in numeric or a in ("--format", "--duration"):
if i + 1 >= len(argv):
raise UsageError(f"{a} needs a value")
v = argv[i + 1]
if a == "--format":
if v not in ("text", "json", "youtube"):
raise UsageError("--format must be text, json, or youtube")
opts["format"] = v
elif a == "--duration":
if not re.fullmatch(r"(?:\d{1,2}:)?\d{1,2}:\d{2}", v):
raise UsageError("--duration must look like 52:30 or 1:05:00")
opts["duration"] = to_sec(v)
else:
key, typ = numeric[a]
try:
opts[key] = typ(v)
except ValueError:
raise UsageError(f"{a} needs a number, got {v!r}")
if opts[key] < 0 or (key in ("wpm", "max_chapters") and opts[key] == 0):
raise UsageError(f"{a} must be positive")
i += 2
elif a.startswith("--"):
raise UsageError(f"unknown option {a}")
elif opts["file"] is None:
opts["file"] = a
i += 1
else:
raise UsageError(f"unexpected argument {a!r}")
if opts["file"] is None:
raise UsageError("give a transcript file, or - for stdin")
return opts
def main(argv):
try:
opts = parse_args(argv)
if opts["file"] == "-":
text = sys.stdin.read()
else:
with open(opts["file"], encoding="utf-8") as f:
text = f.read()
except UsageError as e:
print(f"error: {e}", file=sys.stderr)
return 2
except (OSError, UnicodeDecodeError) as e:
print(f"error: cannot read {opts['file']}: {getattr(e, 'strerror', None) or e}", file=sys.stderr)
return 2
units = parse(text)
total_words = sum(len(words(u["text"])) for u in units if u["kind"] == "text")
if total_words < 50:
print(f"error: only {total_words} words of transcript; need at least 50 to suggest chapters", file=sys.stderr)
return 1
timed = assign_times(units, opts["wpm"], opts["offset"])
last = units[-1]
end_time = last["time"] + (len(words(last["text"])) / opts["wpm"] * 60 if last["kind"] == "text" else 0)
end_time = max(end_time, LAST_CUE_END[0] + opts["offset"] if LAST_CUE_END[0] else 0)
if opts["duration"]:
end_time = max(end_time, opts["duration"])
EXTRA_STOP.update(w.lower() for u in units if u["speaker"] for w in u["speaker"].split())
chosen, _scores = boundaries(units, opts["min_chapter"] * 60, opts["max_chapters"], end_time)
chapters = build(units, chosen, end_time)
quotes = quote_candidates(units, opts["quotes"])
speakers = Counter(u["speaker"] for u in units if u["speaker"])
notes = []
if not timed:
notes.append(f"no timestamps found: times are estimates at {opts['wpm']:g} words per minute; replace them "
"with real times from the audio editor before publishing")
if opts["duration"] and opts["duration"] < end_time - 30 and timed:
notes.append("the transcript runs past --duration; check the timestamps or the offset")
markers = [{"time": round(u["time"]), "time_hms": fmt(u["time"]), "marker": u["text"]}
for u in units if u["kind"] == "pause"]
yt_lines, yt_problems = youtube_lines(chapters)
report = {"markers": markers, "words": total_words, "timestamps": "found" if timed else "estimated", "end": round(end_time),
"end_hms": fmt(end_time), "speakers": dict(speakers.most_common()), "chapters": chapters,
"quotes": quotes, "youtube": yt_lines, "youtube_problems": yt_problems, "notes": notes}
if opts["format"] == "json":
print(json.dumps(report, indent=2, ensure_ascii=False))
return 0
if opts["format"] == "youtube":
print("\n".join(yt_lines))
for p in yt_problems + notes:
print(f"# note: {p}")
return 0
spk = ", ".join(f"{s} ({c} turns)" for s, c in speakers.most_common()) or "not labeled"
print(f"Transcript: {total_words} words, about {fmt(end_time)} long, timestamps {report['timestamps']}, "
f"speakers: {spk}")
print(f"\nChapter candidates ({len(chapters)}, min {opts['min_chapter']:g} min apart):")
for c in chapters:
est = "~" if c["estimated_time"] else " "
print(f" {c['n']:>2}. {est}{c['start_hms']:>8} {c['minutes']:>5.1f} min [{c['source']}] {c['title_draft']}")
print(f" keywords: {', '.join(c['keywords']) or '-'}")
if c["opening_line"]:
print(f" opens: \"{c['opening_line']}\"")
if markers:
print("\nMarkers (breaks, music, ads): " + ", ".join(f"{m['time_hms']} {m['marker']}" for m in markers))
print(f"\nQuote candidates ({len(quotes)}):")
for q in quotes:
who = f"{q['speaker']}: " if q["speaker"] else ""
edited = " (fillers removed, check against audio)" if q["edited"] else ""
print(f" {fmt(q['time']):>8} {who}\"{q['text']}\"{edited}")
print("\nYouTube chapter lines:")
for line in yt_lines:
print(f" {line}")
for p in yt_problems + notes:
print(f"note: {p}")
return 0
if __name__ == "__main__":
sys.exit(main(sys.argv[1:]))Reviews a schema and slow SQL queries, proposes composite indexes in the right column order (equality, sort, range) with covering columns and tradeoffs, flags non-sargable filters like functions on columns and leading-wildcard LIKE, and warns about duplicate, redundant, and low-selectivity indexes. Use when a user shares a slow query, EXPLAIN output, or schema and asks "what index do I need?" or "too many indexes?". Includes a tested stdlib Python helper.
---
name: sql-index-recommendation-reviewer
description: Reviews a schema and slow SQL queries, proposes composite indexes in the right column order (equality, sort, range) with covering columns and tradeoffs, flags non-sargable filters like functions on columns and leading-wildcard LIKE, and warns about duplicate, redundant, and low-selectivity indexes. Use when a user shares a slow query, EXPLAIN output, or schema and asks "what index do I need?" or "too many indexes?". Includes a tested stdlib Python helper.
---
# SQL Index Recommendation Reviewer
You help developers make slow queries fast without burying their tables in indexes. You read the schema and the query the way a database engineer would: which columns are compared with equality, which with ranges, what the query sorts by, what it joins on, and which indexes already exist. You propose the smallest set of indexes that serves the workload, in the right column order, and you are honest that only EXPLAIN on real data can confirm it.
## Files in this skill
- `scripts/suggest_indexes.py` - parses CREATE TABLE / CREATE INDEX statements and SELECT, UPDATE, and DELETE queries (subqueries and CTE bodies separately), then prints candidate indexes with DDL, the queries each one serves, covering columns, overlaps, and schema findings (Python 3 standard library only; no database connection)
- `references/indexing-rules.md` - column order, covering indexes, partial and expression indexes, sargability, and over-indexing rules
- `references/explain-reading-guide.md` - how to read EXPLAIN in PostgreSQL, MySQL, and SQLite to confirm or reject a suggestion
- `templates/index-review.md` - the review report to return
- `examples/example-shop-orders-review.md` - a worked review of an online shop's slow queries
## Workflow
### 1. Collect the facts
Ask, or assume and say so:
- The engine and version (PostgreSQL, MySQL/MariaDB, SQLite, SQL Server), and the schema with existing indexes (`\d table`, `SHOW CREATE TABLE`, or the migration files).
- The slow queries exactly as the app sends them, with typical parameter values, how often each runs, and how slow it is.
- Table sizes (rows), how write-heavy each table is, and any EXPLAIN or EXPLAIN ANALYZE output.
- Constraints: can the team build indexes online, disk space, and replication lag concerns.
### 2. Run the helper
```bash
python3 scripts/suggest_indexes.py --schema schema.sql slow-queries.sql
python3 scripts/suggest_indexes.py all-in-one.sql --dialect mysql
python3 scripts/suggest_indexes.py --schema schema.sql - --json < query.sql
python3 scripts/suggest_indexes.py --schema schema.sql queries.sql --strict # exit 1 on non-sargable filters or duplicate indexes
```
The helper is a heuristic parser, not a planner: it does not know row counts, data distribution, or which plan the database will choose. It can miss predicates in complex SQL (it lists them as "not understood"). Read every query yourself and treat each suggestion as a candidate to test. If you cannot run the script, apply `references/indexing-rules.md` by hand in the same order.
### 3. Decide
1. Fix the query before adding an index: rewrite non-sargable filters (functions on columns, arithmetic, leading-wildcard LIKE, OR across columns) so an index can be used at all.
2. For each query, choose column order: equality columns (most selective first when you know it), then ORDER BY or GROUP BY columns if the query sorts with LIMIT, then one range column.
3. Merge: one composite index can serve several queries when they share its leftmost columns. Prefer extending an existing index over adding a new one.
4. Add covering (INCLUDE) columns only for hot, read-heavy queries, and keep them few.
5. Remove before adding: drop duplicate and redundant-prefix indexes after checking usage statistics (`pg_stat_user_indexes`, `sys.schema_unused_indexes`).
6. Count the cost: every index slows inserts and updates of its columns and uses disk and memory. Write-heavy tables deserve fewer, broader indexes.
### 4. Deliver
Fill in `templates/index-review.md`: per-query diagnosis, the recommended DDL in build order with the queries each serves, indexes to drop, query rewrites, tradeoffs, and how to verify with EXPLAIN before and after (see `references/explain-reading-guide.md`). Follow the style of `examples/example-shop-orders-review.md`.
## Rules
- Never claim a speed-up without saying how to measure it; give expected plan changes ("index scan instead of sequential scan and sort"), not invented timings.
- Use safe build commands for production: `CREATE INDEX CONCURRENTLY` in PostgreSQL, online DDL (`ALGORITHM=INPLACE, LOCK=NONE`) in MySQL, and a maintenance window for SQLite.
- Do not recommend an index on a low-selectivity column alone (status, booleans); use it inside a composite or as a partial index condition.
- Keep primary keys and unique constraints; never suggest dropping them to remove "redundancy".
- Tell the user to build one index at a time and compare plans, and to remove indexes that EXPLAIN never uses.
FILE:references/indexing-rules.md
# Indexing rules
The rules the helper applies, plus the judgment it cannot make. Examples use PostgreSQL syntax unless noted.
## 1. Column order in a composite index
A B-tree index is sorted by its first column, then the second, and so on. A query can seek on a leftmost run of columns compared with equality, then use one more column for a range or for sorting.
Order: **equality columns, then sort columns, then one range column**.
```sql
-- WHERE customer_id = ? AND status = 'paid' ORDER BY created_at DESC LIMIT 20
CREATE INDEX ON orders (customer_id, status, created_at);
```
- Among equality columns, put the one used by the most queries first so the index serves them all; selectivity matters less than reuse once all of them are equalities.
- After a range column (`>`, `<`, `BETWEEN`, `LIKE 'abc%'`), later columns cannot be used for seeking, only for filtering inside the index.
- A sort can use the index only if every ORDER BY column comes from the index in the same order and directions are all the same (or all reversed). Mixed directions need a matching index definition (`created_at DESC, id ASC`).
- With a range and an ORDER BY on a different column, the database must choose: seek with the range and sort, or walk the sort order and filter. With LIMIT, walking the sort order often wins.
## 2. Leftmost prefix and merging
An index on `(a, b, c)` serves queries on `a`, on `a, b`, and on `a, b, c`, but not on `b` or `c` alone. Two candidates where one is a leftmost prefix of the other become one index. An existing index `(a)` is redundant once `(a, b)` exists, unless it is unique, partial, or much smaller and very hot.
## 3. Joins and foreign keys
- Index the column on the side that is looked up: usually the foreign key (`orders.customer_id`). The primary key side is already indexed.
- PostgreSQL does not index foreign keys automatically (MySQL InnoDB does). Missing FK indexes also make deletes on the parent table slow.
- A join column plus that table's own filters can form one composite: `(customer_id, created_at)`.
## 4. Covering indexes
If the index holds every column the query reads, the database can skip the table (index-only scan).
- PostgreSQL 11+ and SQL Server: `INCLUDE (col, ...)` stores extra columns without sorting by them.
- MySQL and SQLite: append the columns to the key; InnoDB secondary indexes already carry the primary key.
- Use it for hot, read-mostly queries with few selected columns. Wide covering indexes are big and slow to write.
- PostgreSQL index-only scans also need a well-vacuumed table (visibility map).
## 5. Partial and expression indexes
- Partial (PostgreSQL, SQLite; filtered index in SQL Server): `CREATE INDEX ON orders (created_at) WHERE status = 'paid' AND shipped_at IS NULL;` Tiny and fast when the condition matches few rows. The query must repeat the same condition.
- Expression: `CREATE INDEX ON users (lower(email));` serves `WHERE lower(email) = ?`. MySQL 8 supports functional indexes `((lower(email)))` or a generated column with an index.
## 6. Sargability: filters an index cannot use
| Pattern | Problem | Fix |
|---|---|---|
| `WHERE date(created_at) = '2026-10-12'` | function on the column | `created_at >= '2026-10-12' AND created_at < '2026-10-13'` |
| `WHERE YEAR(updated_at) = 2026` | function on the column | range on the column |
| `WHERE price * 1.2 > 100` | arithmetic on the column | `price > 100 / 1.2` |
| `WHERE name LIKE '%lamp%'` | leading wildcard | full-text search, or `pg_trgm` GIN index |
| `WHERE lower(name) = ?` | function | expression index on `lower(name)` or a case-insensitive collation |
| `WHERE a = 1 OR b = 2` | OR across columns | `UNION` of two indexed queries, or one index per column (bitmap OR in PostgreSQL) |
| `WHERE status <> 'done'`, `NOT IN` | negative filter | positive list `IN (...)` or a partial index |
| `WHERE id = '42'` on an integer column | implicit cast (MySQL may skip the index on string columns compared to numbers) | compare with the right type |
## 7. Over-indexing
- Every index is updated on INSERT, on DELETE, and on UPDATE of its columns; it also uses disk, memory (buffer cache), and backup and replication bandwidth.
- Duplicates (same columns, same condition) are pure cost: drop one.
- Single-column indexes on low-selectivity columns (booleans, status, type) are rarely used; the planner prefers a scan when a value matches more than about 5 to 20 percent of rows. Use them inside composites or as partial-index conditions.
- As a rule of thumb, more than 5 or 6 indexes on a write-heavy table deserves a review; read-mostly tables can carry more.
- Find unused indexes before dropping: PostgreSQL `pg_stat_user_indexes.idx_scan = 0` (since the last stats reset, on every replica), MySQL `sys.schema_unused_indexes`. Drop with `DROP INDEX CONCURRENTLY` in PostgreSQL.
## 8. Building safely
- PostgreSQL: `CREATE INDEX CONCURRENTLY` avoids blocking writes; it cannot run inside a transaction and leaves an INVALID index on failure (drop and retry).
- MySQL 8: `ALTER TABLE ... ADD INDEX ..., ALGORITHM=INPLACE, LOCK=NONE`; watch replication lag on large tables.
- SQLite: index builds lock the database; run during low traffic.
- Build one index at a time, check the plan, then the next.
FILE:references/explain-reading-guide.md
# Reading EXPLAIN to confirm an index
Always compare the plan before and after, on production-like data and with realistic parameter values. Small test tables make the planner choose scans that are right for small tables.
## PostgreSQL
```sql
EXPLAIN (ANALYZE, BUFFERS) SELECT ...; -- runs the query; wrap writes in BEGIN ... ROLLBACK
```
| Plan node | Meaning |
|---|---|
| Seq Scan | reads the whole table; fine for small tables or when most rows match |
| Index Scan | seeks the index, then reads matching table rows |
| Index Only Scan | answered from the index; check "Heap Fetches" is low (vacuum if not) |
| Bitmap Index Scan + Bitmap Heap Scan | collects many matches, then reads pages in order; common for medium selectivity and OR |
| Sort (with "external merge") | sorting spilled to disk; an index in ORDER BY order can remove it |
| Rows Removed by Filter | rows read and thrown away; high numbers mean the index does not cover the filter |
Look at: actual time of the top node, rows estimated versus actual (a 10x gap means stale statistics: run `ANALYZE table`), and shared hit/read buffers. Confirm a new index is used with `pg_stat_user_indexes.idx_scan` after a day of traffic.
## MySQL 8
```sql
EXPLAIN FORMAT=TREE SELECT ...;
EXPLAIN ANALYZE SELECT ...; -- 8.0.18+, runs the query
```
In classic `EXPLAIN` output check: `type` (ALL = full scan, ref/eq_ref/range = index use, index = full index scan), `key` (the index chosen), `rows` (estimate), and `Extra` ("Using index" = covering, "Using filesort" = extra sort, "Using temporary" = temporary table for GROUP BY or DISTINCT). Refresh statistics with `ANALYZE TABLE t`.
## SQLite
```sql
EXPLAIN QUERY PLAN SELECT ...;
```
"SCAN t" is a full scan; "SEARCH t USING INDEX idx (a=? AND b>?)" shows which columns seek; "USING COVERING INDEX" means no table access; "USE TEMP B-TREE FOR ORDER BY" means an extra sort. Run `ANALYZE;` so the planner has statistics.
## What success looks like
- The slow query's plan switches from a full scan plus sort to an index (only) scan with no sort, and the rows read drop close to the rows returned.
- Writes on the table do not get noticeably slower (check p95 insert and update latency after deploying).
- No other important query changed plan for the worse; run the top queries by total time (`pg_stat_statements`, MySQL performance schema digest tables) before and after.
## When the index is not used
1. Statistics are stale: run ANALYZE.
2. The filter is not sargable (see indexing-rules.md section 6).
3. The value matches too many rows; the planner is right to scan.
4. Column types or collations differ between the query and the column (implicit cast).
5. A parameterized query uses a generic plan; test with the real prepared statement.
6. The index columns are in the wrong order for the query.
FILE:templates/index-review.md
# Index review: app_or_service
**Engine:** engine_and_version | **Tables:** tables_with_row_counts | **Write load:** write_profile
**Helper:** n_queries queries analyzed, n_suggested candidates, n_findings schema findings
## Summary
Two or three sentences: the main cause of slowness, the indexes to add and drop, and the expected plan change.
## Per-query diagnosis
| Query | Runs | Now (plan / time) | Problem | Fix |
|---|---|---|---|---|
| Q1 | per_minute | Seq Scan + Sort | | |
## Query rewrites (do these first)
```sql
-- before
-- after
```
## Indexes to add (build in this order)
| # | DDL | Serves | Why this column order | Cost |
|---|---|---|---|---|
| 1 | | | | |
## Indexes to drop (after checking usage)
| Index | Reason | Check first |
|---|---|---|
| | duplicate / redundant prefix / unused / low selectivity | `idx_scan` on primary and replicas |
## Net effect per table
| Table | Indexes before | After | Note |
|---|---|---|---|
## How to verify
1. Capture `EXPLAIN (ANALYZE, BUFFERS)` for each query before.
2. Build one index (`CONCURRENTLY` / online DDL), run `ANALYZE`, capture the plan again.
3. Expect: (plan change per query).
4. Watch write latency and replication lag for a day; then drop the indexes listed above.
## Caveats
- Suggestions come from query text and schema only; row counts and data skew can change the best choice.
FILE:examples/example-shop-orders-review.md
# Example: index review for an online shop
## User request
"Our PostgreSQL 16 shop database is getting slow. orders has 12 million rows and gets about 40 inserts a second, customers has 900k rows, products 20k. Here are the schema and the six queries that show up in pg_stat_statements. What indexes should we add, and do we have too many already?"
## Helper run
```
$ python3 scripts/suggest_indexes.py --schema shop-schema.sql slow-queries.sql
Schema: 4 table(s), 12 existing index(es) incl. keys | dialect: postgres
Queries analyzed: 7
Q1 SELECT: SELECT o.id, o.total, o.created_at FROM orders o WHERE o.customer_id = $1 AND o.status = 'paid' ORDER BY o.cre...
equality: orders.customer_id (=), orders.status (=)
order by: orders.created_at DESC (LIMIT)
Q2 SELECT: SELECT c.email, count(*) AS orders FROM customers c JOIN orders o ON o.customer_id = c.id WHERE c.country = 'D...
equality: customers.country (=)
range: orders.created_at (range)
joins: orders.customer_id = customers.id
group by: customers.email
Q3 SELECT: SELECT * FROM products WHERE lower(name) LIKE '%lamp%'
WARN non-sargable: lower(name) in the WHERE clause hides the column from a normal index; ... the LIKE pattern '%lamp%' also starts with a wildcard, so only full-text search or a trigram index can help
Q4 SELECT: ... WHERE oi.order_id IN (SELECT id FROM orders WHERE customer_id = 42 AND created_at > now() - interval '30 days')
note: subquery analyzed as Q5
Q6 UPDATE: UPDATE orders SET status = 'shipped', shipped_at = now() WHERE status = 'paid' AND shipped_at IS NULL AND crea...
Q7 SELECT: SELECT id FROM customers WHERE email = 'ann@example.com' OR phone = '+491701234567'
WARN OR across different columns (...): one composite index cannot serve both sides; consider UNION ALL of two indexed queries or one index per side
Suggested indexes (4):
1. CREATE INDEX CONCURRENTLY idx_customers_country_email ON customers (country, email);
2. CREATE INDEX CONCURRENTLY idx_orders_customer_id_created_at ON orders (customer_id, created_at);
serves: Q2, Q5 | overlap: same leading column as #3
3. CREATE INDEX CONCURRENTLY idx_orders_customer_id_status_created_at ON orders (customer_id, status, created_at) INCLUDE (id, total);
serves: Q1 | overlap: same leading column as #2
4. CREATE INDEX CONCURRENTLY idx_orders_status_shipped_at_created_at ON orders (status, shipped_at, created_at);
serves: Q6 | alternative: ... ON orders (status, created_at) WHERE shipped_at IS NULL;
Schema findings (5):
WARN redundant-prefix customers: idx_customers_country (country) is a leftmost prefix of idx_customers_country_created ...
WARN low-selectivity customers: idx_customers_active indexes only 'is_active' ...
HIGH duplicate-index orders: idx_orders_created and orders_created_at_idx have the same columns (created_at); drop one
WARN low-selectivity orders: idx_orders_status indexes only 'status' ...
WARN many-indexes orders: 4 existing + 3 suggested = 7 indexes (limit 6) ...
```
## Review (filled template, shortened)
**Summary:** orders has no index that starts with customer_id, so the order history page (Q1) and the recent-orders lookup (Q5) scan millions of rows, and two of its four indexes are a duplicate and a low-selectivity status index. Add one composite and one partial index on orders, a phone index and a trigram index elsewhere, and drop three indexes: orders ends with the same count of 4 but every one is useful.
| Query | Problem | Fix |
|---|---|---|
| Q1 order history | no index on customer_id; sort on created_at | (customer_id, created_at) with status and total included: newest rows for the customer, status filtered inside the index, LIMIT stops after 20 |
| Q2 country report | join needs orders.customer_id; customers.country is already served by idx_customers_country_created | same orders index; no new customers index (the helper's #1 only helps the GROUP BY, and the report runs once a night) |
| Q3 product search | `lower(name) LIKE '%lamp%'` cannot use a B-tree | pg_trgm GIN index on lower(name) |
| Q4/Q5 recent items | subquery needs (customer_id, created_at); order_items PK (order_id, product_id) serves the IN | same orders index |
| Q6 shipping job | scans by status | partial index on created_at for paid, unshipped orders only |
| Q7 support lookup | OR across email and phone | UNION of two lookups; email is unique already, add a phone index |
Why not the helper's #3 as well: #2 and #3 share customer_id, and on a table with 40 inserts a second one index is better than two. The user confirmed that about 95 percent of orders are paid, so Q1 throws away few rows, and (customer_id, created_at) with status in INCLUDE keeps Q1 fast. If the share of non-paid orders grows, switch to #3's column order.
**Indexes to add (in this order):**
```sql
CREATE INDEX CONCURRENTLY idx_orders_customer_created
ON orders (customer_id, created_at) INCLUDE (id, status, total); -- Q1, Q2, Q5
CREATE INDEX CONCURRENTLY idx_orders_paid_unshipped
ON orders (created_at) WHERE status = 'paid' AND shipped_at IS NULL; -- Q6, holds only paid, unshipped orders
CREATE INDEX CONCURRENTLY idx_customers_phone ON customers (phone); -- Q7 after the rewrite
CREATE EXTENSION IF NOT EXISTS pg_trgm;
CREATE INDEX CONCURRENTLY idx_products_name_trgm
ON products USING gin (lower(name) gin_trgm_ops); -- Q3, query unchanged
```
**Query rewrite (Q7):**
```sql
SELECT id FROM customers WHERE email = 'ann@example.com'
UNION
SELECT id FROM customers WHERE phone = '+491701234567';
```
**Indexes to drop (after checking `pg_stat_user_indexes.idx_scan` on the primary and both replicas):**
| Index | Reason |
|---|---|
| orders_created_at_idx | duplicate of idx_orders_created |
| idx_orders_status | low selectivity; Q6 now uses the partial index |
| idx_customers_country | leftmost prefix of idx_customers_country_created |
| idx_customers_active (check) | boolean; keep only if idx_scan shows real use |
**Net effect:** orders 4 -> 4 indexes (two replaced by better ones), customers 5 -> 5 (one dropped, phone added; 4 if is_active goes too), products 2 -> 3.
**Recheck with the helper** (schema after the change, queries after the rewrite; query numbers follow the new file, so Q3 is the old Q5 and Q4/Q5 are the two UNION branches):
```
$ python3 scripts/suggest_indexes.py --schema shop-schema-after.sql shop-queries-after.sql --no-covering
Already served by existing indexes (5):
customers (email) -> customers_email_key (Q4)
customers (id, country) -> customers_pkey (Q2)
customers (phone) -> idx_customers_phone (Q5)
orders (created_at) -> idx_orders_created (Q2)
orders (customer_id, created_at) -> idx_orders_customer_created (Q2, Q3)
```
The helper still lists (customer_id, status, created_at) for Q1 and (country, email) for Q2; both were rejected on purpose for the reasons above. There are no duplicate or redundant-prefix findings left.
**How to verify:** before and after each build, run `EXPLAIN (ANALYZE, BUFFERS)` for Q1 with a real customer id. Expect "Index Scan using idx_orders_customer_created" (or Index Only Scan) with no Sort node and about 20 rows read instead of a Bitmap Heap Scan over thousands. Watch insert latency on orders for a day before dropping the old indexes.
FILE:scripts/suggest_indexes.py
#!/usr/bin/env python3
"""Suggest candidate indexes for slow SQL queries from a schema and the query text (no database needed).
Usage:
python3 suggest_indexes.py --schema schema.sql queries.sql [options]
python3 suggest_indexes.py all-in-one.sql (CREATE statements and queries in one file)
python3 suggest_indexes.py --schema schema.sql - < slow.sql
Options:
--schema FILE CREATE TABLE / CREATE INDEX statements (may also be inside the query file)
--dialect NAME postgres (default), mysql, or sqlite; changes INCLUDE and online-build hints
--max-indexes N warn when a table would have more than N indexes (default 6)
--no-covering do not suggest covering (INCLUDE) columns
--json print the result as JSON
--strict exit 1 if any query has a non-sargable predicate or the schema has duplicate indexes
What it does: lightweight parsing of SELECT, UPDATE, and DELETE statements (subqueries are analyzed
separately) to find equality filters, range filters, join columns, ORDER BY and GROUP BY columns,
and selected columns per table. It builds composite candidates (equality columns, then sort, then
range), merges candidates that are prefixes of each other, skips those already served by an existing
index, and reports duplicate and redundant indexes, low-selectivity single-column indexes, and tables
with many indexes. It does not know row counts or data distribution: confirm every suggestion with
EXPLAIN on real data.
Exit codes: 0 ok, 1 --strict findings, 2 usage or read error.
Python 3 standard library only.
"""
import json
import re
import signal
import sys
if hasattr(signal, "SIGPIPE"):
signal.signal(signal.SIGPIPE, signal.SIG_DFL) # quiet exit when piped into head
KEYWORDS = set("""select from where and or not in is null like ilike between join inner left right full outer cross on
using group by order having limit offset as asc desc distinct union all exists case when then else end update set
delete insert into values returning with true false interval current_date current_timestamp now lateral natural
fetch first rows only for share nowait skip locked""".split())
LOW_CARD = re.compile(r"^(is_|has_|can_|should_)|^(status|state|type|kind|flag|active|enabled|deleted|archived|gender|"
r"visible|published|verified|priority|level|tier|role)$", re.I)
CLAUSE_END = r"(?=\b(?:where|group\s+by|order\s+by|having|limit|offset|union|returning|for\s+update|window)\b|$)"
class UsageError(Exception):
pass
# ---------------------------------------------------------------- text helpers
def strip_comments(sql):
sql = re.sub(r"/\*.*?\*/", " ", sql, flags=re.S)
return re.sub(r"--[^\n]*", " ", sql)
def split_statements(sql):
out, buf, quote = [], [], None
for ch in sql:
if quote:
buf.append(ch)
if ch == quote:
quote = None
continue
if ch in "'\"`":
quote = ch
if ch == ";":
out.append("".join(buf))
buf = []
else:
buf.append(ch)
out.append("".join(buf))
return [s.strip() for s in out if s.strip()]
def unquote_idents(sql):
sql = re.sub(r'"([A-Za-z_][\w$]*)"', r"\1", sql)
sql = re.sub(r"`([A-Za-z_][\w$]*)`", r"\1", sql)
return re.sub(r"\[([A-Za-z_][\w$]*)\]", r"\1", sql)
def mask_strings(sql):
"""Replace string literals with placeholders, remembering LIKE patterns."""
lits = []
def rep(m):
lits.append(m.group(0)[1:-1])
return f" __lit{len(lits) - 1}__ "
return re.sub(r"'(?:[^']|'')*'", rep, sql), lits
def split_top(text, sep_regex):
"""Split on a keyword at parenthesis depth 0."""
parts, depth, last = [], 0, 0
for m in re.finditer(r"\(|\)|" + sep_regex, text, flags=re.I):
tok = m.group(0)
if tok == "(":
depth += 1
elif tok == ")":
depth -= 1
elif depth == 0:
parts.append(text[last:m.start()])
last = m.end()
parts.append(text[last:])
return [p.strip() for p in parts if p.strip()]
def split_and(text):
"""Split on top-level AND, keeping 'x BETWEEN a AND b' together."""
parts = split_top(text, r"\band\b")
out = []
for part in parts:
if out and re.search(r"\bbetween\s+[^\s]+$", out[-1], re.I):
out[-1] = f"{out[-1]} AND {part}"
else:
out.append(part)
return out
def extract_subqueries(sql):
"""Pull out innermost (SELECT ...) blocks; return (outer_sql, [subqueries])."""
subs = []
pattern = re.compile(r"\(\s*(select\b[^()]*(?:\([^()]*\)[^()]*)*)\)", re.I)
while True:
m = pattern.search(sql)
if not m:
break
subs.append(m.group(1))
sql = sql[:m.start()] + " __subq__ " + sql[m.end():]
return sql, subs
# ---------------------------------------------------------------- schema
def parse_columns_list(text):
return [re.sub(r"\s+(asc|desc)\b.*$", "", c.strip(), flags=re.I).strip().lower()
for c in split_top(text, r",")]
def parse_schema(statements):
tables, indexes = {}, []
for st in statements:
s = unquote_idents(strip_comments(st)).strip()
m = re.match(r"create\s+table\s+(?:if\s+not\s+exists\s+)?([\w.]+)\s*\((.*)\)\s*[^)]*$", s, re.I | re.S)
if m:
name = m.group(1).split(".")[-1].lower()
cols, pk = [], None
for part in split_top(m.group(2), r","):
p = part.strip()
low = p.lower()
c = re.match(r"(?:constraint\s+\w+\s+)?primary\s+key\s*\(([^)]*)\)", low)
if c:
pk = parse_columns_list(c.group(1))
continue
c = re.match(r"(?:constraint\s+(\w+)\s+)?unique\s*(?:key\s+\w+\s*)?\(([^)]*)\)", low)
if c:
indexes.append({"name": c.group(1) or f"{name}_unique", "table": name,
"columns": parse_columns_list(c.group(2)), "unique": True, "kind": "unique"})
continue
c = re.match(r"(?:key|index)\s+(\w+)\s*\(([^)]*)\)", low)
if c:
indexes.append({"name": c.group(1), "table": name, "columns": parse_columns_list(c.group(2)),
"unique": False, "kind": "index"})
continue
if re.match(r"(constraint|foreign\s+key|check|exclude)\b", low):
continue
col = re.match(r"([a-z_][\w$]*)\s+", low + " ")
if col:
cols.append(col.group(1))
if re.search(r"\bprimary\s+key\b", low):
pk = [col.group(1)]
elif re.search(r"\bunique\b", low):
indexes.append({"name": f"{name}_{col.group(1)}_key", "table": name,
"columns": [col.group(1)], "unique": True, "kind": "unique"})
tables[name] = {"columns": cols}
if pk:
indexes.append({"name": f"{name}_pkey", "table": name, "columns": pk, "unique": True,
"kind": "primary"})
continue
m = re.match(r"create\s+(unique\s+)?index\s+(?:concurrently\s+)?(?:if\s+not\s+exists\s+)?(\w+)\s+on\s+"
r"([\w.]+)\s*(?:using\s+\w+\s*)?\(([^;]*?)\)\s*(include\s*\(([^)]*)\))?\s*(where\s+.*)?$",
s, re.I | re.S)
if m:
cols = parse_columns_list(m.group(4))
indexes.append({"name": m.group(2), "table": m.group(3).split(".")[-1].lower(), "columns": cols,
"unique": bool(m.group(1)), "kind": "index",
"include": parse_columns_list(m.group(6)) if m.group(6) else [],
"partial": (m.group(7) or "").strip() or None})
return tables, indexes
# ---------------------------------------------------------------- queries
def parse_from(from_text):
"""Return (tables {alias: table}, order list, join predicates text list)."""
aliases, order, ons = {}, [], []
pieces = re.split(r"\b((?:natural\s+)?(?:inner\s+|left\s+(?:outer\s+)?|right\s+(?:outer\s+)?|full\s+(?:outer\s+)?|"
r"cross\s+)?join)\b", from_text, flags=re.I)
items = [pieces[0]] + [pieces[i + 1] for i in range(1, len(pieces) - 1, 2)]
for item in items:
on_split = re.split(r"\bon\b", item, maxsplit=1, flags=re.I)
using = re.search(r"\busing\s*\(([^)]*)\)", on_split[0], re.I)
head = re.sub(r"\busing\s*\([^)]*\)", "", on_split[0], flags=re.I)
for ref in head.split(","):
toks = [t for t in re.split(r"\s+", ref.strip()) if t and t.lower() not in ("as", "lateral", "only")]
if not toks or toks[0] == "__subq__":
continue
table = toks[0].split(".")[-1].lower()
alias = toks[1].lower() if len(toks) > 1 else table
aliases[alias] = table
order.append(alias)
if len(on_split) > 1:
ons.append(on_split[1])
if using and len(order) >= 2:
for col in parse_columns_list(using.group(1)):
ons.append(f"{order[-2]}.{col} = {order[-1]}.{col}")
return aliases, order, ons
COL = r"((?:[a-z_][\w$]*\.)?[a-z_][\w$]*)"
VALUE = r"(?:__lit\d+__|\?|:\w+|\$\d+|@\w+|%s|-?\d+(?:\.\d+)?|true|false|null|__subq__|current_date|current_timestamp|now\(\))"
def resolve(ref, aliases, tables):
ref = ref.lower()
if "." in ref:
a, c = ref.split(".", 1)
return (aliases.get(a, a), c)
if len(set(aliases.values())) == 1:
return (next(iter(aliases.values())), ref)
owners = [t for t in set(aliases.values()) if ref in tables.get(t, {}).get("columns", [])]
return (owners[0], ref) if len(owners) == 1 else (None, ref)
def analyze_predicate(pred, lits, aliases, tables, q):
p = pred.strip()
while p.startswith("(") and p.endswith(")") and len(split_top(p[1:-1], r"\band\b")) >= 1 and \
p.count("(") == p.count(")") and _balanced_wrap(p):
p = p[1:-1].strip()
if len(split_and(p)) > 1:
for sub in split_and(p):
analyze_predicate(sub, lits, aliases, tables, q)
return
ors = split_top(p, r"\bor\b")
if len(ors) > 1:
cols = set()
for o in ors:
m = re.match(rf"^\(?\s*{COL}\s*(=|in\b)", o.strip(), re.I)
cols.add(m.group(1).lower() if m else None)
if len(cols) == 1 and None not in cols:
t, c = resolve(cols.pop(), aliases, tables)
q["filters"].append({"table": t, "column": c, "kind": "eq", "op": "OR of equalities (same as IN)"})
else:
shown = re.sub(r"__lit(\d+)__", lambda m: "'" + lits[int(m.group(1))] + "'", p)
shown = shown if len(shown) <= 70 else shown[:67] + "..."
q["warnings"].append(f"OR across different columns ({shown}): one composite index cannot serve "
"both sides; consider UNION ALL of two indexed queries or one index per side")
return
if re.match(r"^not\b", p, re.I) or re.search(r"(<>|!=|\bnot\s+in\b|\bnot\s+like\b)", p, re.I):
m = re.match(rf"^(?:not\s+)?{COL}", p, re.I)
shown = re.sub(r"__lit(\d+)__", lambda m: "'" + lits[int(m.group(1))] + "'", p)[:50]
q["notes"].append(f"negative predicate '{shown}' rarely uses an index; it is applied as a filter")
return
m = re.match(rf"^{COL}\s*(=|<=>)\s*{COL}$", p, re.I)
if m and not re.fullmatch(VALUE, m.group(3), re.I) and m.group(3).lower() not in KEYWORDS:
left, right = resolve(m.group(1), aliases, tables), resolve(m.group(3), aliases, tables)
if left[0] and right[0] and left[0] != right[0]:
q["joins"].append((left, right))
return
m = re.match(rf"^{VALUE}\s*(=|<|>|<=|>=)\s*{COL}$", p, re.I)
if m:
flip = {"<": ">", ">": "<", "<=": ">=", ">=": "<=", "=": "="}
p = f"{m.group(2)} {flip[m.group(1)]} __v__"
m = re.match(rf"^{COL}\s*(=|<=>)\s*(.+)$", p, re.I)
if m and m.group(1).lower() not in KEYWORDS:
t, c = resolve(m.group(1), aliases, tables)
q["filters"].append({"table": t, "column": c, "kind": "eq", "op": "="})
return
m = re.match(rf"^{COL}\s+in\s+__subq__$", p, re.I)
if m:
t, c = resolve(m.group(1), aliases, tables)
q["filters"].append({"table": t, "column": c, "kind": "eq", "op": "IN (subquery)"})
return
m = re.match(rf"^{COL}\s+in\s*\((.*)\)$", p, re.I | re.S)
if m:
t, c = resolve(m.group(1), aliases, tables)
q["filters"].append({"table": t, "column": c, "kind": "eq", "op": "IN"})
return
m = re.match(rf"^{COL}\s+is\s+(not\s+)?null$", p, re.I)
if m:
t, c = resolve(m.group(1), aliases, tables)
if m.group(2):
q["notes"].append(f"'{c} IS NOT NULL' is usually not selective; a partial index can help if most rows are NULL")
else:
q["filters"].append({"table": t, "column": c, "kind": "eq", "op": "IS NULL"})
return
m = re.match(rf"^{COL}\s*(<|>|<=|>=)\s*(.+)$", p, re.I) or re.match(rf"^{COL}\s+between\s+(.+)\s+and\s+(.+)$", p, re.I)
if m and m.group(1).lower() not in KEYWORDS:
t, c = resolve(m.group(1), aliases, tables)
q["filters"].append({"table": t, "column": c, "kind": "range", "op": "range"})
return
m = re.match(rf"^{COL}\s+(i?like)\s+__lit(\d+)__", p, re.I)
if m:
t, c = resolve(m.group(1), aliases, tables)
pattern = lits[int(m.group(3))]
if pattern.startswith(("%", "_")):
q["warnings"].append(f"non-sargable: {c} {m.group(2).upper()} '{pattern}' starts with a wildcard, so a "
"B-tree index cannot seek; use full-text search or a trigram index (PostgreSQL pg_trgm)")
q["non_sargable"] = True
else:
q["filters"].append({"table": t, "column": c, "kind": "range", "op": f"LIKE '{pattern}'"})
if m.group(2).lower() == "ilike":
q["notes"].append(f"ILIKE on {c}: a plain B-tree index is case-sensitive; index lower({c}) instead")
return
m = re.match(rf"^([a-z_]+)\s*\(\s*{COL}[^)]*\)\s*(=|<|>|<=|>=|like|between)", p, re.I)
if m and m.group(1).lower() not in KEYWORDS:
func, col = m.group(1).lower(), m.group(2)
t, c = resolve(col, aliases, tables)
wild = re.search(r"like\s+__lit(\d+)__", p, re.I)
extra = ""
if wild and lits[int(wild.group(1))].startswith(("%", "_")):
extra = (f"; the LIKE pattern '{lits[int(wild.group(1))]}' also starts with a wildcard, so only full-text "
"search or a trigram index can help")
q["warnings"].append(f"non-sargable: {func}({c}) in the WHERE clause hides the column from a normal index; "
f"rewrite as a range on {c} or add an expression index on {func}({c}){extra}")
q["non_sargable"] = True
return
m = re.match(rf"^{COL}\s*[-+*/]\s*\S+\s*(=|<|>|<=|>=)", p, re.I)
if m:
t, c = resolve(m.group(1), aliases, tables)
q["warnings"].append(f"non-sargable: arithmetic on {c} ('{p[:40]}'); move the arithmetic to the other side")
q["non_sargable"] = True
return
if "__subq__" in p and re.search(r"\bexists\b", p, re.I):
return
q["notes"].append(f"predicate not understood, check by hand: {p[:60]}")
def _balanced_wrap(p):
depth = 0
for i, ch in enumerate(p):
depth += ch == "("
depth -= ch == ")"
if depth == 0 and i < len(p) - 1:
return False
return True
def parse_query(raw, tables, number, lits=None, outer_aliases=None):
if lits is None:
sql, lits = mask_strings(unquote_idents(strip_comments(raw)))
else:
sql = raw
sql = re.sub(r"\s+", " ", sql).strip()
outer, subs = extract_subqueries(sql)
shown = re.sub(r"__lit(\d+)__", lambda m: "'" + lits[int(m.group(1))] + "'", re.sub(r"\s+", " ", raw.strip()))
shown = re.sub(r"\s+", " ", shown)
q = {"n": number, "text": shown[:110] + ("..." if len(shown) > 110 else ""), "kind": None, "aliases": {},
"filters": [], "joins": [], "order_by": [], "group_by": [], "selected": {}, "select_star": False,
"limit": False, "warnings": [], "notes": [], "non_sargable": False, "subqueries": subs, "lits": lits}
low = outer.lower()
cte_names = set()
if low.startswith("with "):
cte_names = {n.lower() for n in re.findall(r"(?:with|,)\s*(?:recursive\s+)?(\w+)\s+as\b", outer, re.I)}
q["notes"].append(f"CTE {', '.join(sorted(cte_names))} analyzed as separate queries; the final SELECT reads from it")
m = re.search(r"(?:\)|__subq__)\s*((?:select|update|delete)\b.*)$", outer, re.I)
outer = m.group(1) if m else outer
low = outer.lower()
if low.startswith("select"):
q["kind"] = "SELECT"
m = re.match(r"select\s+(?:distinct\s+)?(.*?)\s+from\s+(.*?)" + CLAUSE_END, outer, re.I)
if not m:
q["notes"].append("no FROM clause found")
return q
select_list, from_text = m.group(1), m.group(2)
elif low.startswith("update"):
q["kind"] = "UPDATE"
m = re.match(r"update\s+(.*?)\s+set\s+", outer, re.I)
select_list, from_text = "", m.group(1) if m else ""
elif low.startswith("delete"):
q["kind"] = "DELETE"
m = re.match(r"delete\s+from\s+(.*?)" + CLAUSE_END, outer, re.I)
select_list, from_text = "", m.group(1) if m else ""
else:
return None
own, _order, ons = parse_from(from_text)
own = {a: t for a, t in own.items() if t not in cte_names}
q["aliases"] = own
aliases = {**(outer_aliases or {}), **own} # outer aliases resolve correlated references
unknown = [t for t in set(own.values()) if tables and t not in tables]
if unknown:
q["notes"].append(f"table(s) not in the schema: {', '.join(sorted(unknown))}; existing indexes unknown")
for on in ons:
for pred in split_and(on):
analyze_predicate(pred, lits, aliases, tables, q)
w = re.search(r"\bwhere\b(.*?)(?=\b(?:group\s+by|order\s+by|having|limit|offset|returning|for\s+update)\b|$)", outer, re.I)
if w:
for pred in split_and(w.group(1)):
analyze_predicate(pred, lits, aliases, tables, q)
for clause, key in (("group\\s+by", "group_by"), ("order\\s+by", "order_by")):
m = re.search(rf"\b{clause}\b(.*?)(?=\b(?:having|order\s+by|limit|offset|for\s+update)\b|$)", outer, re.I)
if m:
for item in split_top(m.group(1), r","):
mm = re.match(rf"^{COL}\s*(asc|desc)?(?:\s+nulls\s+(?:first|last))?$", item.strip(), re.I)
if mm:
t, c = resolve(mm.group(1), aliases, tables)
q[key].append({"table": t, "column": c, "dir": (mm.group(2) or "asc").lower()})
else:
q["notes"].append(f"{key.replace('_', ' ')} expression '{item.strip()[:40]}' cannot use a plain column index")
q["limit"] = bool(re.search(r"\b(limit|fetch\s+first)\b", outer, re.I))
if select_list:
if re.search(r"(^|,)\s*(\w+\.)?\*\s*(,|$)", select_list):
q["select_star"] = True
for ref in re.findall(rf"\b{COL}\b(?!\s*\()", select_list, re.I):
if ref.lower() in KEYWORDS or ref.startswith("__") or re.fullmatch(r"\d+", ref):
continue
t, c = resolve(ref, aliases, tables)
if t and (not tables or c in tables.get(t, {}).get("columns", [c])):
q["selected"].setdefault(t, [])
if c not in q["selected"][t]:
q["selected"][t].append(c)
for f in q["filters"]:
if f["table"] is None:
q["notes"].append(f"column '{f['column']}' could belong to several tables; qualify it with an alias")
return q
# ---------------------------------------------------------------- candidates
def candidates_for(q, dialect, covering, pk_cols=None):
out = []
by_table = {}
for f in q["filters"]:
if f["table"]:
by_table.setdefault(f["table"], {"eq": [], "range": []})
lst = by_table[f["table"]][f["kind"]]
if f["column"] not in lst:
lst.append(f["column"])
tables_in_q = set(q["aliases"].values())
correlated = set()
for left, right in q["joins"]:
if left[0] in tables_in_q and right[0] not in tables_in_q:
correlated.add(left[0])
if right[0] in tables_in_q and left[0] not in tables_in_q:
correlated.add(right[0])
order_tables = {o["table"] for o in q["order_by"]}
for t in tables_in_q:
if t in correlated:
continue # reached through the correlated join below
eq = by_table.get(t, {}).get("eq", [])
rng = [c for c in by_table.get(t, {}).get("range", []) if c not in eq]
sort = []
if q["order_by"] and order_tables == {t} and len({o["dir"] for o in q["order_by"]}) == 1:
sort = [o["column"] for o in q["order_by"] if o["column"] not in eq]
elif q["group_by"] and {g["table"] for g in q["group_by"]} == {t}:
sort = [g["column"] for g in q["group_by"] if g["column"] not in eq]
key, why = list(eq), []
if eq:
why.append(f"equality on {', '.join(eq)} first")
if sort and (not rng or q["limit"]):
key += [c for c in sort if c not in key]
purpose = "already sorted" if q["order_by"] else "already grouped"
why.append(f"then {', '.join(sort)} so rows come out {purpose}" + (" (stops early with LIMIT)" if q["limit"] else ""))
if rng and rng[0] not in key:
key.append(rng[0])
why.append(f"{rng[0]} last as an in-index filter")
elif rng:
key.append(rng[0])
why.append(f"then the range on {rng[0]}" if eq else f"range on {rng[0]}")
if len(rng) > 1:
q["notes"].append(f"several range filters on {t} ({', '.join(rng)}): only the first can seek; the rest filter")
if sort:
q["notes"].append(f"ORDER BY on {t} cannot use the same index as the range on {rng[0]}; expect a sort step")
if not key:
continue
include = []
if covering and not q["select_star"] and q["kind"] == "SELECT":
extra = [c for c in q["selected"].get(t, []) if c not in key
and not (dialect == "mysql" and c in (pk_cols or {}).get(t, []))]
if 0 < len(extra) <= 3:
include = extra
out.append({"table": t, "columns": key, "include": include, "why": "; ".join(why), "queries": [q["n"]],
"source": "filter", "eq": list(eq),
"null_filter": [f["column"] for f in q["filters"] if f["table"] == t and f["op"] == "IS NULL"]})
for left, right in q["joins"]:
for (t, c), (ot, oc) in ((left, right), (right, left)):
if t not in tables_in_q:
continue # outer table of a correlated subquery
eq = [e for e in by_table.get(t, {}).get("eq", []) if e != c]
rng = [r for r in by_table.get(t, {}).get("range", []) if r != c and r not in eq][:1]
why = f"join column {t}.{c} = {ot}.{oc} first (lets the planner look up matching rows)"
if eq or rng:
why += f", then {', '.join(eq + rng)} to filter inside the index"
if t in correlated:
why = f"correlated lookup {t}.{c} = outer {ot}.{oc} first" + (f", then {', '.join(eq + rng)}" if eq or rng else "")
out.append({"table": t, "columns": [c] + eq + rng, "include": [], "queries": [q["n"]], "source": "join",
"eq": [c] + eq, "why": why})
return out
def served_by(cand, indexes):
for ix in indexes:
if ix["table"] != cand["table"]:
continue
cols = ix["columns"]
if cand["source"] == "join" and ix["unique"] and cols == cand["columns"][:1] and not ix.get("partial"):
return ix # joining on a unique key: one row per lookup, extra columns add little
if cols[:len(cand["columns"])] == cand["columns"]:
return ix
n_eq = len(cand.get("eq", []))
if n_eq > 1 and len(cols) >= len(cand["columns"]) and set(cols[:n_eq]) == set(cand["columns"][:n_eq]) \
and cols[n_eq:len(cand["columns"])] == cand["columns"][n_eq:]:
return ix # same equality columns in another order serve the query equally well
return None
def served_by_named(c, indexes):
ix = served_by(c, indexes)
return {"existing": ix["name"] if ix else "?"}
def merge(cands):
merged = []
for c in sorted(cands, key=lambda c: -len(c["columns"])):
for m in merged:
if m["table"] == c["table"] and m["columns"][:len(c["columns"])] == c["columns"]:
m["queries"] = sorted(set(m["queries"]) | set(c["queries"]))
m["include"] = m["include"] or []
break
else:
merged.append(dict(c))
return sorted(merged, key=lambda c: (c["table"], c["columns"]))
def schema_findings(indexes, max_indexes, new_by_table):
out = []
by_table = {}
for ix in indexes:
by_table.setdefault(ix["table"], []).append(ix)
for t, ixs in by_table.items():
for i, a in enumerate(ixs):
for b in ixs[i + 1:]:
if a["columns"] == b["columns"] and (a.get("partial") == b.get("partial")):
out.append({"level": "high", "rule": "duplicate-index",
"message": f"{t}: {a['name']} and {b['name']} have the same columns ({', '.join(a['columns'])}); drop one"})
elif a["columns"] != b["columns"]:
short, long_ = (a, b) if len(a["columns"]) < len(b["columns"]) else (b, a)
if long_["columns"][:len(short["columns"])] == short["columns"] and not short["unique"] \
and not short.get("partial"):
out.append({"level": "warn", "rule": "redundant-prefix",
"message": f"{t}: {short['name']} ({', '.join(short['columns'])}) is a leftmost prefix of "
f"{long_['name']} ({', '.join(long_['columns'])}); usually droppable"})
for ix in ixs:
if len(ix["columns"]) == 1 and LOW_CARD.search(ix["columns"][0]) and ix["kind"] == "index" and not ix.get("partial"):
out.append({"level": "warn", "rule": "low-selectivity",
"message": f"{t}: {ix['name']} indexes only '{ix['columns'][0]}', which sounds like a flag or status "
"with few values; use a composite or partial index instead"})
total = len(ixs) + new_by_table.get(t, 0)
if total > max_indexes:
out.append({"level": "warn", "rule": "many-indexes",
"message": f"{t}: {len(ixs)} existing + {new_by_table.get(t, 0)} suggested = {total} indexes "
f"(limit {max_indexes}); every insert and update maintains all of them"})
return out
def ddl(c, dialect):
name = f"idx_{c['table']}_{'_'.join(c['columns'])}"[:63]
cols = ", ".join(c["columns"])
if c["include"]:
if dialect == "postgres":
return f"CREATE INDEX CONCURRENTLY {name} ON {c['table']} ({cols}) INCLUDE ({', '.join(c['include'])});"
cols += ", " + ", ".join(c["include"])
if dialect == "postgres":
return f"CREATE INDEX CONCURRENTLY {name} ON {c['table']} ({cols});"
if dialect == "mysql":
return f"ALTER TABLE {c['table']} ADD INDEX {name} ({cols}), ALGORITHM=INPLACE, LOCK=NONE;"
return f"CREATE INDEX {name} ON {c['table']} ({cols});"
# ---------------------------------------------------------------- main
def parse_args(argv):
opts = {"schema": None, "files": [], "dialect": "postgres", "max_indexes": 6, "covering": True,
"json": False, "strict": False}
i = 0
while i < len(argv):
a = argv[i]
if a in ("-h", "--help"):
print(__doc__)
sys.exit(0)
if a in ("--schema", "--dialect", "--max-indexes"):
if i + 1 >= len(argv):
raise UsageError(f"{a} needs a value")
v = argv[i + 1]
if a == "--schema":
opts["schema"] = v
elif a == "--dialect":
if v not in ("postgres", "mysql", "sqlite"):
raise UsageError("--dialect must be postgres, mysql, or sqlite")
opts["dialect"] = v
else:
try:
opts["max_indexes"] = int(v)
except ValueError:
raise UsageError(f"--max-indexes needs a whole number, got {v!r}")
i += 2
elif a == "--no-covering":
opts["covering"] = False; i += 1
elif a == "--json":
opts["json"] = True; i += 1
elif a == "--strict":
opts["strict"] = True; i += 1
elif a.startswith("--"):
raise UsageError(f"unknown option {a}")
else:
opts["files"].append(a); i += 1
if not opts["files"] and not opts["schema"]:
raise UsageError("give a query file (or - for stdin), optionally with --schema FILE")
return opts
def read(path):
if path == "-":
return sys.stdin.read()
with open(path, encoding="utf-8") as f:
return f.read()
def main(argv):
try:
opts = parse_args(argv)
schema_text = read(opts["schema"]) if opts["schema"] else ""
query_texts = [read(p) for p in opts["files"]]
except UsageError as e:
print(f"error: {e}", file=sys.stderr)
return 2
except OSError as e:
print(f"error: cannot read {e.filename}: {e.strerror}", file=sys.stderr)
return 2
statements = split_statements(strip_comments(schema_text))
query_stmts = []
for text in query_texts:
clean = strip_comments(text)
for st in split_statements(clean):
(statements if re.match(r"\s*create\s", st, re.I) else query_stmts).append(st)
tables, indexes = parse_schema(statements)
queries, n = [], 0
expanded = []
for st in query_stmts:
masked, _l = mask_strings(st)
if len(split_top(masked, r"\bunion(?:\s+all)?\b")) > 1: # a real top-level UNION, not one inside a string
expanded.extend(split_top(st, r"\bunion(?:\s+all)?\b")) # each branch is analyzed as its own query
else:
expanded.append(st)
for st in expanded:
pending = [(st, None, None)]
while pending:
cur, lits, outer = pending.pop(0)
n += 1
q = parse_query(cur, tables, n, lits, outer)
if q is None:
n -= 1
continue
queries.append(q)
for sub in q["subqueries"]:
pending.append((sub, q["lits"], {**(outer or {}), **q["aliases"]}))
q["notes"].append(f"subquery analyzed as Q{n + len(pending)}")
if not queries:
print("error: no SELECT, UPDATE, or DELETE statements found", file=sys.stderr)
return 2
pks = {ix["table"]: ix["columns"] for ix in indexes if ix["kind"] == "primary"}
cands = []
for q in queries:
cands += candidates_for(q, opts["dialect"], opts["covering"], pks)
covered, open_cands = [], []
for c in cands:
ix = served_by(c, indexes)
if ix:
covered.append({**c, "existing": ix["name"]})
else:
open_cands.append(c)
suggested = merge(open_cands)
covered = merge(covered)
for c in covered:
c["existing"] = next(x["existing"] for x in [served_by_named(c, indexes)])
new_by_table = {}
for c in suggested:
new_by_table[c["table"]] = new_by_table.get(c["table"], 0) + 1
c["ddl"] = ddl(c, opts["dialect"])
lead = c["columns"][0]
c["tradeoff"] = ("every INSERT, UPDATE of these columns, and DELETE on " + c["table"] +
" also writes this index; size grows with the row count and column widths")
if LOW_CARD.search(lead) and len(c["columns"]) == 1:
c["tradeoff"] += f"; '{lead}' looks low-selectivity, so the planner may still prefer a scan"
for a in suggested:
others = [k for k, b in enumerate(suggested, 1) if b is not a and b["table"] == a["table"]
and b["columns"][0] == a["columns"][0]]
a["overlaps"] = others
a["makes_redundant"] = [ix["name"] for ix in indexes if ix["table"] == a["table"] and not ix["unique"]
and ix["kind"] != "primary" and len(ix["columns"]) < len(a["columns"])
and a["columns"][:len(ix["columns"])] == ix["columns"]]
if a.get("null_filter") and opts["dialect"] == "postgres":
col = a["null_filter"][0]
rest = [c for c in a["columns"] if c != col]
name = f"idx_{a['table']}_{'_'.join(rest)}_{col}_null"[:63]
a["partial_hint"] = (f"CREATE INDEX CONCURRENTLY {name} ON {a['table']} ({', '.join(rest)}) "
f"WHERE {col} IS NULL; -- smaller if few rows match")
findings = schema_findings(indexes, opts["max_indexes"], new_by_table)
report = {"dialect": opts["dialect"], "tables": {t: len(v["columns"]) for t, v in tables.items()},
"existing_indexes": indexes, "queries": [{k: v for k, v in q.items() if k not in ("subqueries", "lits")}
for q in queries],
"suggested": suggested, "already_covered": covered, "schema_findings": findings}
for q in report["queries"]:
q["joins"] = [f"{a[0]}.{a[1]} = {b[0]}.{b[1]}" for a, b in q["joins"]]
strict_fail = opts["strict"] and (any(q["non_sargable"] for q in queries) or
any(f["rule"] == "duplicate-index" for f in findings))
if opts["json"]:
print(json.dumps(report, indent=2))
return 1 if strict_fail else 0
print(f"Schema: {len(tables)} table(s), {len(indexes)} existing index(es) incl. keys | dialect: {opts['dialect']}")
print(f"Queries analyzed: {len(queries)}")
for q in report["queries"]:
print(f"\nQ{q['n']} {q['kind']}: {q['text']}")
tabs = ", ".join(f"{t} {a}" if a != t else t for a, t in q["aliases"].items())
print(f" tables: {tabs or '-'}")
eqs = [f"{f['table'] or '?'}.{f['column']} ({f['op']})" for f in q["filters"] if f["kind"] == "eq"]
rngs = [f"{f['table'] or '?'}.{f['column']} ({f['op']})" for f in q["filters"] if f["kind"] == "range"]
if eqs:
print(f" equality: {', '.join(eqs)}")
if rngs:
print(f" range: {', '.join(rngs)}")
if q["joins"]:
print(f" joins: {', '.join(q['joins'])}")
if q["order_by"]:
print(" order by: " + ", ".join(f"{o['table']}.{o['column']} {o['dir'].upper()}" for o in q["order_by"])
+ (" (LIMIT)" if q["limit"] else ""))
if q["group_by"]:
print(" group by: " + ", ".join(f"{g['table']}.{g['column']}" for g in q["group_by"]))
for w in q["warnings"]:
print(f" WARN {w}")
for nt in q["notes"]:
print(f" note: {nt}")
print(f"\nSuggested indexes ({len(suggested)}):")
if not suggested:
print(" none: existing indexes already lead with the needed columns")
for k, c in enumerate(suggested, 1):
print(f" {k}. {c['ddl']}")
print(f" serves: {', '.join('Q' + str(x) for x in c['queries'])} | why: {c['why']}")
if c["include"]:
print(f" covering: {', '.join(c['include'])} added so the query can skip the table (optional)")
print(f" tradeoff: {c['tradeoff']}")
if c.get("overlaps"):
print(f" overlap: same leading column as #{', #'.join(map(str, c['overlaps']))}; "
"check whether one index can serve both before creating two")
if c.get("makes_redundant"):
print(f" makes redundant: {', '.join(c['makes_redundant'])} (leftmost prefix of this index; drop after checking usage)")
if c.get("partial_hint"):
print(f" alternative: {c['partial_hint']}")
if covered:
print(f"\nAlready served by existing indexes ({len(covered)}):")
for c in covered:
print(f" {c['table']} ({', '.join(c['columns'])}) -> {c['existing']} (Q{', Q'.join(map(str, c['queries']))})")
print(f"\nSchema findings ({len(findings)}):")
if not findings:
print(" none")
for f in findings:
print(f" {f['level'].upper():<4} {f['rule']:<17} {f['message']}")
print("\nNext: run EXPLAIN (ANALYZE, BUFFERS) or EXPLAIN on real data before and after; build one index at a time.")
return 1 if strict_fail else 0
if __name__ == "__main__":
sys.exit(main(sys.argv[1:]))Lints .env and .env.example files for secrets committed by mistake, weak names, duplicate keys, export prefixes, spaces around equals, empty required values, and unsafe interpolation, then suggests a cleaned example file. Use when a user asks "is this .env safe to commit?", "why won't dotenv load?", or "diff my .env.example". Includes a tested stdlib Python linter.
---
name: dotenv-config-linter
description: Lints .env and .env.example files for secrets committed by mistake, weak names, duplicate keys, export prefixes, spaces around equals, empty required values, and unsafe interpolation, then suggests a cleaned example file. Use when a user asks "is this .env safe to commit?", "why won't dotenv load?", or "diff my .env.example". Includes a tested stdlib Python linter.
---
# Dotenv Config and Secrets Linter
You help people keep environment files correct and safe. You distinguish secrets that must never be committed from placeholders that belong in `.env.example`, catch parse problems that break `dotenv` loaders, and propose a cleaned example file. You never ask the user to paste production secrets into chat if a redacted sample will do.
## Files in this skill
- `scripts/dotenv_lint.py` - parses dotenv syntax, reports HIGH/WARN/INFO findings, and can emit a sanitized `.env.example` (Python 3 standard library only)
- `references/dotenv-rules.md` - syntax and safety rules the linter encodes
- `references/secret-patterns.md` - common secret shapes and how to replace them with placeholders
- `templates/lint-report.md` - the report shape to return
- `examples/example-webapp-env-review.md` - a worked review of a messy web app env file
## Workflow
### 1. Collect the facts
- Whether the file is `.env` (local secrets), `.env.example` (placeholders), or CI env config.
- Which keys are required at boot.
- Whether the repo already has a secret scanner (gitleaks, trufflehog).
### 2. Run the linter
```bash
python3 scripts/dotenv_lint.py .env
python3 scripts/dotenv_lint.py .env.example --example-mode
python3 scripts/dotenv_lint.py app.env --emit-example cleaned.env.example
python3 scripts/dotenv_lint.py - --json < sample.env
```
Prefer redacted samples. If the user pastes real secrets, warn them to rotate anything that may have leaked.
### 3. Fix
1. Treat HIGH findings (likely real secrets in an example file, duplicate keys, unquoted values with spaces) first.
2. Align `.env.example` with required keys using placeholders from `references/secret-patterns.md`.
3. Deliver with `templates/lint-report.md`.
## Rules
- Never recommend committing a real `.env` with live credentials.
- Do not print full secret values in summaries; show masked forms (`sk-…XXXX`).
- Keep key names stable unless the user wants a rename for clarity.
FILE:references/dotenv-rules.md
# Dotenv rules this skill enforces
## Syntax
- One `KEY=VALUE` per line; `#` starts a comment.
- Prefer no spaces around `=`. `KEY = VALUE` is WARN (some loaders accept it, many do not).
- Values with spaces or `#` should be single- or double-quoted.
- `export KEY=VALUE` is INFO (works in shells, redundant for most dotenv parsers).
- Duplicate keys: last one wins in many loaders — flag as HIGH because intent is unclear.
## Safety
- `.env.example` must not contain values that look like live secrets (long random tokens, private keys, `sk-`/`ghp_` prefixes, AWS-shaped keys).
- Empty values for keys listed as required → WARN.
- Interpolation `OTHER` is INFO; confirm the referenced key exists.
## Naming
- Prefer `SCREAMING_SNAKE_CASE`.
- Avoid generic names like `PASSWORD` without a service prefix (`DB_PASSWORD`).
FILE:references/secret-patterns.md
# Secret patterns → placeholders
| Looks like | Placeholder |
|---|---|
| `sk-` + long alnum | `sk-your-openai-key` |
| `ghp_` / `github_pat_` | `ghp_your_token` |
| `AKIA` + 16 chars | `AKIA_YOUR_ACCESS_KEY` |
| `-----BEGIN` private key block | omit from example; document path instead |
| Long base64-ish 32+ chars | `replace-me` or `changeme_<purpose>` |
| Database URLs with passwords | `postgres://USER:PASSWORD@host:5432/db` |
When emitting an example file, replace secret-looking values with placeholders and keep non-secret config (ports, feature flags, public URLs) as-is if they are clearly non-sensitive.
FILE:templates/lint-report.md
# Dotenv lint report
## File reviewed
- Path / role: `.env` | `.env.example` | other
## Summary
- HIGH: …
- WARN: …
- INFO: …
## Findings
| Severity | Line | Key | Issue | Fix |
|---|---|---|---|---|
## Suggested `.env.example`
```dotenv
…
```
## Rotate if exposed
- …
FILE:examples/example-webapp-env-review.md
# Example: messy webapp `.env.example`
**Input (bad):**
```
export DB_URL=postgres://app:SuperSecret123@db:5432/app
API_KEY = sk-proj-abcdefghijklmnopqrstuvwxyz012345
PORT=3000
PORT=3001
FEATURE_X=true
```
**Findings:**
- HIGH: `API_KEY` looks like a live OpenAI-style key in an example file.
- HIGH: duplicate `PORT`.
- WARN: spaces around `=` on `API_KEY`.
- INFO: `export` prefix on `DB_URL`.
- HIGH: database password embedded in example URL.
**Cleaned example:**
```
DB_URL=postgres://USER:PASSWORD@db:5432/app
API_KEY=sk-your-openai-key
PORT=3000
FEATURE_X=true
```
FILE:scripts/dotenv_lint.py
#!/usr/bin/env python3
"""Lint .env / .env.example files for syntax and likely secrets.
Usage:
python3 dotenv_lint.py .env
python3 dotenv_lint.py .env.example --example-mode
python3 dotenv_lint.py app.env --emit-example cleaned.env.example
python3 dotenv_lint.py - --json < sample.env
"""
from __future__ import annotations
import argparse
import json
import re
import sys
from pathlib import Path
SECRET_RES = [
(re.compile(r"sk-[A-Za-z0-9-]{20,}"), "openai-like key"),
(re.compile(r"ghp_[A-Za-z0-9]{20,}"), "github token"),
(re.compile(r"github_pat_[A-Za-z0-9_]{20,}"), "github pat"),
(re.compile(r"AKIA[0-9A-Z]{16}"), "aws access key id"),
(re.compile(r"-----BEGIN [A-Z ]*PRIVATE KEY-----"), "private key block"),
(re.compile(r"(?i)password=([^&\s]{8,})"), "password in URL"),
(re.compile(r"(?i)://[^:/@]+:[^@/]{8,}@"), "credentials embedded in URL"),
]
LINE_RE = re.compile(
r"^(?P<export>export\s+)?(?P<key>[A-Za-z_][A-Za-z0-9_]*)\s*=\s*(?P<val>.*)$"
)
def mask(val: str) -> str:
if len(val) <= 8:
return "***"
return val[:3] + "…" + val[-4:]
def looks_placeholder(val: str) -> bool:
v = val.strip().strip("'\"")
low = v.lower()
if (
not v
or "your-" in low
or "changeme" in low
or "replace" in low
or "example" in low
or low in {"password", "user", "todo", "xxx"}
or "user:password@" in low
or "://user:password@" in low
):
return True
# URL with obvious placeholders
if "://" in v and "@" in v:
if re.search(r"://(USER|user|USERNAME|username):(PASSWORD|password|PASS|pass|secret)@", v):
return True
return False
def parse_file(text: str) -> tuple[list[dict], list[dict]]:
findings = []
entries = []
seen = {}
for i, line in enumerate(text.splitlines(), 1):
raw = line.rstrip("\n")
stripped = raw.strip()
if not stripped or stripped.startswith("#"):
continue
m = LINE_RE.match(stripped)
if not m:
findings.append({"severity": "HIGH", "line": i, "key": None,
"issue": "unparseable line", "fix": "use KEY=VALUE"})
continue
key, val, export = m.group("key"), m.group("val"), m.group("export")
if re.search(r"\s=\s|=\s|\s=", raw.split("#")[0]) and " = " in raw:
findings.append({"severity": "WARN", "line": i, "key": key,
"issue": "spaces around equals", "fix": f"{key}={val.strip()}"})
if export:
findings.append({"severity": "INFO", "line": i, "key": key,
"issue": "export prefix", "fix": "drop export for dotenv parsers"})
if key in seen:
findings.append({"severity": "HIGH", "line": i, "key": key,
"issue": f"duplicate key (also line {seen[key]})",
"fix": "keep a single definition"})
seen[key] = i
# strip inline comments for unquoted values
store_val = val
if val and not (val.startswith("'") or val.startswith('"')):
if " #" in val:
store_val = val.split(" #", 1)[0].rstrip()
entries.append({"line": i, "key": key, "value": store_val, "raw": raw})
return findings, entries
def secret_findings(entries: list[dict], example_mode: bool) -> list[dict]:
out = []
for e in entries:
val = e["value"].strip().strip("'\"")
if looks_placeholder(val):
continue
for rx, label in SECRET_RES:
if rx.search(val):
sev = "HIGH" if example_mode else "WARN"
out.append({"severity": sev, "line": e["line"], "key": e["key"],
"issue": f"likely secret ({label}): {mask(val)}",
"fix": "replace with placeholder in example files; rotate if committed"})
break
else:
# long random-ish token
if example_mode and len(val) >= 32 and re.fullmatch(r"[A-Za-z0-9+/=_-]+", val):
out.append({"severity": "HIGH", "line": e["line"], "key": e["key"],
"issue": f"long token-like value: {mask(val)}",
"fix": "use a placeholder like replace-me"})
return out
def emit_example(entries: list[dict]) -> str:
lines = []
seen = set()
for e in entries:
if e["key"] in seen:
continue
seen.add(e["key"])
val = e["value"].strip().strip("'\"")
secret = False
if not looks_placeholder(val):
for rx, _ in SECRET_RES:
if rx.search(val):
secret = True
break
if len(val) >= 32 and re.fullmatch(r"[A-Za-z0-9+/=_-]+", val):
secret = True
if secret or (val and not looks_placeholder(val) and re.search(r"(?i)secret|password|token|api[_-]?key", e["key"])):
if re.search(r"(?i)api[_-]?key|token", e["key"]):
new_v = "replace-me-token"
elif re.search(r"(?i)password|secret", e["key"]):
new_v = "changeme"
elif "://" in val and "@" in val:
new_v = re.sub(r"://([^:/@]+):([^@/]+)@", r"://USER:PASSWORD@", val)
else:
new_v = "replace-me"
else:
new_v = val
lines.append(f"{e['key']}={new_v}")
return "\n".join(lines) + ("\n" if lines else "")
def main(argv: list[str]) -> int:
ap = argparse.ArgumentParser(description=__doc__)
ap.add_argument("path", help="Env file path, or - for stdin")
ap.add_argument("--example-mode", action="store_true",
help="Treat file as .env.example (secrets are HIGH)")
ap.add_argument("--emit-example", metavar="PATH", help="Write sanitized example file")
ap.add_argument("--json", action="store_true")
args = ap.parse_args(argv)
if args.path == "-":
text = sys.stdin.read()
else:
path = Path(args.path)
if not path.is_file():
print(f"error: file not found: {args.path}", file=sys.stderr)
return 2
text = path.read_text(encoding="utf-8")
example_mode = args.example_mode or (args.path != "-" and Path(args.path).name.endswith(".example"))
findings, entries = parse_file(text)
findings.extend(secret_findings(entries, example_mode))
findings.sort(key=lambda f: ({"HIGH": 0, "WARN": 1, "INFO": 2}[f["severity"]], f["line"] or 0))
example = emit_example(entries)
if args.emit_example:
Path(args.emit_example).write_text(example, encoding="utf-8")
summary = {"HIGH": 0, "WARN": 0, "INFO": 0}
for f in findings:
summary[f["severity"]] += 1
if args.json:
json.dump({"summary": summary, "findings": findings, "example": example}, sys.stdout, indent=2)
sys.stdout.write("\n")
else:
print(f"summary: HIGH={summary['HIGH']} WARN={summary['WARN']} INFO={summary['INFO']}")
for f in findings:
key = f["key"] or "-"
print(f"[{f['severity']}] line {f['line']}: {key}: {f['issue']} → {f['fix']}")
if args.emit_example:
print(f"wrote example: {args.emit_example}")
return 1 if summary["HIGH"] else 0
if __name__ == "__main__":
sys.exit(main(sys.argv[1:]))Counts syllables per line with a transparent heuristic, checks common forms (haiku, limerick, iambic pentameter targets), maps end-rhyme schemes (AABB, ABAB, ABCB), and flags near-misses with suggested edits. Use when a user asks "does this scan?", "is this a haiku?", or "what is my rhyme scheme?". Includes a tested stdlib Python checker.
---
name: poetry-meter-and-rhyme-checker
description: Counts syllables per line with a transparent heuristic, checks common forms (haiku, limerick, iambic pentameter targets), maps end-rhyme schemes (AABB, ABAB, ABCB), and flags near-misses with suggested edits. Use when a user asks "does this scan?", "is this a haiku?", or "what is my rhyme scheme?". Includes a tested stdlib Python checker.
---
# Poetry Meter and Rhyme Checker
You help writers see how a draft scans: syllables per line, a plausible stress sketch for English meter targets, and the end-rhyme scheme. You are honest that English syllable counting is heuristic, you show your work, and you suggest small edits instead of rewriting the poem unless asked.
## Files in this skill
- `scripts/poetry_check.py` - syllable estimates, form checks, rhyme-scheme detection (Python 3 standard library only)
- `references/forms-and-targets.md` - syllable and rhyme targets for common short forms
- `references/syllable-heuristic.md` - how the counter works and where it fails (names, -ed, diphthongs)
- `templates/poetry-report.md` - the report shape to return
- `examples/example-haiku-and-quatrain.md` - worked examples
## Workflow
### 1. Collect the facts
- The poem text (keep the user's line breaks).
- Target form if any: haiku, tanka, limerick, sonnet couplet, free verse with a rhyme check only, or custom syllable pattern.
- Dialect notes (US/UK spelling) when relevant to endings like `-our` / `-or`.
### 2. Run the checker
```bash
python3 scripts/poetry_check.py poem.txt
python3 scripts/poetry_check.py poem.txt --form haiku
python3 scripts/poetry_check.py poem.txt --form limerick --rhyme
python3 scripts/poetry_check.py - --json < poem.txt
```
### 3. Advise
1. Show a per-line syllable table and the detected rhyme scheme.
2. Say whether the form target passed, and which lines miss by how many syllables.
3. Offer minimal edits (swap a synonym, split a compound, drop a filler word) rather than a full rewrite.
4. Deliver with `templates/poetry-report.md`.
## Rules
- Always label syllable counts as estimates.
- Do not mock the writer's voice; keep suggestions optional.
- For free verse, still report syllables and rhyme if useful, without forcing a form.
FILE:references/forms-and-targets.md
# Forms and targets
| Form | Syllable target (lines) | Typical rhyme |
|---|---|---|
| Haiku (English teaching form) | 5 / 7 / 5 | usually unrhymed |
| Tanka | 5 / 7 / 5 / 7 / 7 | usually unrhymed |
| Limerick | ~8–9 / 8–9 / 5–6 / 5–6 / 8–9 | AABBA |
| Quatrain (ballad) | often 8 / 6 / 8 / 6 | ABCB or ABAB |
| Heroic couplet | ~10 / 10 (iambic pentameter feel) | AA |
Exact classical Japanese mora counts differ from English syllable teaching forms; this skill uses the common English classroom targets unless the user asks for mora-aware analysis.
FILE:references/syllable-heuristic.md
# Syllable heuristic (honest limits)
The script estimates English syllables by:
1. Lowercasing and stripping punctuation.
2. Counting vowel groups (`a e i o u y`) with common silent-`e` and `-ed` / `-es` adjustments.
3. Treating hyphenated compounds as the sum of parts.
## Known failure modes
- Proper names (`Olivia`, `Michael`) and loanwords.
- Words where `y` is consonantal (`beyond`) vs vocalic (`rhythm` is still hard).
- `-ism`, `-tion`, and compressed pronunciations (`family` as 2 or 3).
When a count looks wrong, say so and offer a manual override in the report.
FILE:templates/poetry-report.md
# Poetry meter and rhyme report
## Target form
- …
## Per-line table
| # | Line | Syllables (est.) | End word | Rhyme letter |
|---|---|---|---|---|
## Form check
- Pass / fail: …
- Misses: …
## Rhyme scheme
- …
## Minimal edit ideas
- …
## Notes on uncertain counts
- …
FILE:examples/example-haiku-and-quatrain.md
# Examples
## Haiku (should pass 5-7-5)
```
An old silent pond
A frog jumps into the pond—
Splash! Silence again.
```
Estimated syllables often land near 5 / 7 / 5 depending on `silent` and `silence`.
## Quatrain rhyme ABCB
```
The woods are lovely, dark and deep,
But I have promises to keep,
And miles to go before I sleep,
And miles to go before I sleep.
```
End words deep/keep/sleep/sleep → AAAA if all rhyme together; a true ABCB sample should use different B and C words.
FILE:scripts/poetry_check.py
#!/usr/bin/env python3
"""Estimate syllables, check simple forms, and map end-rhyme schemes.
Usage:
python3 poetry_check.py poem.txt
python3 poetry_check.py poem.txt --form haiku
python3 poetry_check.py poem.txt --form limerick --rhyme
python3 poetry_check.py - --json < poem.txt
"""
from __future__ import annotations
import argparse
import json
import re
import sys
from pathlib import Path
FORMS = {
"haiku": [5, 7, 5],
"tanka": [5, 7, 5, 7, 7],
"limerick": None, # soft ranges handled separately
}
VOWELS = set("aeiouy")
def strip_word(w: str) -> str:
return re.sub(r"[^a-zA-Z']", "", w).lower()
def syllables_word(word: str) -> int:
w = strip_word(word)
if not w:
return 0
# common overrides
overrides = {
"poem": 2, "poems": 2, "fire": 1, "hour": 1, "our": 1, "//": 0,
"family": 3, "every": 3, "different": 3, "interesting": 4,
"rhythm": 2, "business": 2, "wednesday": 2, "beautiful": 3,
}
if w in overrides:
return overrides[w]
if "-" in word:
return sum(syllables_word(p) for p in word.split("-"))
# remove trailing silent e
base = w
if base.endswith("e") and not base.endswith(("le", "ue", "oe")) and len(base) > 2:
base = base[:-1]
# -ed endings: walked=1 extra often silent
if re.search(r"[^aeiou]ed$", w) and not w.endswith(("ted", "ded")):
base = re.sub(r"ed$", "", w)
if base.endswith("e"):
base = base[:-1]
groups = re.findall(r"[aeiouy]+", base)
count = len(groups) if groups else 1
# adjust ion/ial
if w.endswith(("tion", "sion", "cial", "tial")):
count = max(count, 2)
return max(1, count)
def end_word(line: str) -> str:
words = [w for w in re.findall(r"[A-Za-z']+", line)]
return words[-1].lower() if words else ""
def rhyme_key(word: str) -> str:
w = strip_word(word)
if not w:
return ""
aliases = {
"you": "oo", "blue": "oo", "too": "oo", "two": "oo", "to": "oo", "through": "oo",
"threw": "oo", "shoe": "oo", "who": "oo", "new": "oo", "knew": "oo", "true": "oo",
"lee": "ee", "tea": "ee", "glee": "ee", "sea": "ee", "me": "ee", "we": "ee", "be": "ee",
"green": "een", "keen": "een", "seen": "een", "mean": "een",
}
if w in aliases:
return aliases[w]
matches = list(re.finditer(r"[aeiouy]+", w))
if not matches:
return w
start = matches[-1].start()
if len(matches) >= 2 and len(w) - start < 2:
start = matches[-2].start()
tail = w[start:]
# light normalization for common English end-rhymes
for a, b in (
("igh", "i"), ("ea", "ee"), ("ee", "ee"), ("ey", "ee"), ("ie", "ee"),
("ay", "ei"), ("ai", "ei"), ("ou", "ow"), ("oo", "u"), ("ue", "u"),
):
if a in tail:
tail = tail.replace(a, b)
return tail
def rhyme_scheme(end_words: list[str]) -> str:
mapping = {}
letters = []
next_ord = 0
for w in end_words:
key = rhyme_key(w)
if not key:
letters.append("-")
continue
# near-rhyme: same key
if key not in mapping:
# also merge if share last 2 chars and vowel
found = None
for k, let in mapping.items():
if k == key or (len(k) >= 2 and len(key) >= 2 and k[-2:] == key[-2:]):
found = let
mapping[key] = let
break
if found is None:
let = chr(ord("A") + next_ord)
next_ord += 1
mapping[key] = let
letters.append(mapping[key])
return "".join(letters)
def limerick_ok(counts: list[int]) -> tuple[bool, str]:
if len(counts) != 5:
return False, f"expected 5 lines, got {len(counts)}"
ranges = [(7, 10), (7, 10), (4, 7), (4, 7), (7, 10)]
misses = []
for i, (c, (lo, hi)) in enumerate(zip(counts, ranges), 1):
if not (lo <= c <= hi):
misses.append(f"line {i}: {c} not in {lo}-{hi}")
return (not misses), ("; ".join(misses) if misses else "ok")
def analyze(text: str, form: str | None) -> dict:
lines = [ln.rstrip() for ln in text.splitlines() if ln.strip()]
rows = []
counts = []
ends = []
for i, ln in enumerate(lines, 1):
words = re.findall(r"[A-Za-z']+(?:-[A-Za-z']+)*", ln)
sc = sum(syllables_word(w) for w in words)
ew = end_word(ln)
counts.append(sc)
ends.append(ew)
rows.append({"n": i, "line": ln, "syllables": sc, "end": ew})
scheme = rhyme_scheme(ends)
form_result = None
if form == "limerick":
ok, detail = limerick_ok(counts)
rhyme_ok = scheme == "AABBA"
form_result = {"form": form, "pass": ok and rhyme_ok, "detail": detail,
"rhyme_expected": "AABBA", "rhyme_got": scheme}
elif form in FORMS and FORMS[form]:
target = FORMS[form]
if len(counts) != len(target):
form_result = {"form": form, "pass": False,
"detail": f"expected {len(target)} lines, got {len(counts)}",
"target": target, "got": counts}
else:
misses = [f"line {i}: got {g}, want {t}" for i, (g, t) in enumerate(zip(counts, target), 1) if g != t]
form_result = {"form": form, "pass": not misses, "detail": "; ".join(misses) or "ok",
"target": target, "got": counts}
return {"lines": rows, "rhyme_scheme": scheme, "form_check": form_result}
def main(argv: list[str]) -> int:
ap = argparse.ArgumentParser(description=__doc__)
ap.add_argument("path", help="Poem file, or - for stdin")
ap.add_argument("--form", choices=["haiku", "tanka", "limerick"], help="Form to validate")
ap.add_argument("--rhyme", action="store_true", help="Always print rhyme scheme (default on)")
ap.add_argument("--json", action="store_true")
args = ap.parse_args(argv)
text = sys.stdin.read() if args.path == "-" else Path(args.path).read_text(encoding="utf-8")
result = analyze(text, args.form)
if args.json:
json.dump(result, sys.stdout, indent=2)
sys.stdout.write("\n")
else:
print("n | syl | end | line")
for r in result["lines"]:
print(f"{r['n']:<2}| {r['syllables']:<3}| {r['end']:<12}| {r['line']}")
print(f"rhyme scheme: {result['rhyme_scheme']}")
fc = result["form_check"]
if fc:
print(f"form {fc['form']}: {'PASS' if fc['pass'] else 'FAIL'} ({fc['detail']})")
if "rhyme_expected" in fc:
print(f" rhyme expected {fc['rhyme_expected']}, got {fc['rhyme_got']}")
if result["form_check"] and not result["form_check"]["pass"]:
return 1
return 0
if __name__ == "__main__":
sys.exit(main(sys.argv[1:]))Compares an OpenAPI YAML/JSON spec to a list of implemented routes and flags drift: missing operations, extra routes, method mismatches, and path-parameter shape differences. Use when reviewing API PRs, before a release, or asking "does our code still match the OpenAPI?". Includes a stdlib Python checker with --self-test.
---
name: openapi-spec-drift-checker
description: Compares an OpenAPI YAML/JSON spec to a list of implemented routes and flags drift: missing operations, extra routes, method mismatches, and path-parameter shape differences. Use when reviewing API PRs, before a release, or asking "does our code still match the OpenAPI?". Includes a stdlib Python checker with --self-test.
---
# OpenAPI Spec Drift Checker
You help API teams find drift between an OpenAPI 3.x document and the routes their code actually serves. You compare paths and methods, flag missing or extra operations, and call out parameter or response mismatches that break clients. You do not invent endpoints; you only report what the inputs show.
## Files in this skill
- `scripts/check_drift.py` - compares a small OpenAPI snippet to a route list and prints a drift report (Python 3 standard library only)
- `references/openapi-checklist.md` - what to verify in the spec before comparing
- `references/common-drift.md` - frequent causes of drift and how to fix them
- `templates/drift-report.md` - report structure for PRs and release notes
- `examples/openapi-snippet.yaml` - sample OpenAPI fragment for demos
## Workflow
### 1. Collect inputs
Ask for, or infer:
- OpenAPI file (YAML or JSON) or a pasted paths section
- Implemented routes: method + path list from the framework router, generated from code, or a gateway config
- Whether path-parameter names must match exactly (`{id}` vs `{userId}`)
### 2. Normalize
Lowercase methods. Strip trailing slashes. Treat `/users/{id}` and `/users/{userId}` as the same path shape unless the team requires exact parameter names.
### 3. Run the checker
```
python3 scripts/check_drift.py --spec examples/openapi-snippet.yaml --routes examples/routes.txt
python3 scripts/check_drift.py --self-test
```
If routes are not in a file yet, help the user extract them from code first.
### 4. Classify findings
- **missing_in_code**: in OpenAPI, not implemented
- **missing_in_spec**: implemented, not in OpenAPI
- **method_mismatch**: path exists but verbs differ
- **param_shape_diff**: same path pattern, different parameter names (optional strict mode)
### 5. Answer
Return a short summary counts table, then each finding with severity (high/medium/low) and a suggested fix. Use `templates/drift-report.md` for PR comments.
## Rules
- Prefer documenting reality when the code is intentional and the spec is stale; prefer implementing when the spec is the contract already published to clients.
- Never claim security review was done; only report structural drift.
- If the OpenAPI cannot be parsed, stop and show the parse error.
FILE:references/openapi-checklist.md
# OpenAPI pre-compare checklist
Before running drift checks, confirm the spec is coherent:
1. **openapi** field is `3.0.x` or `3.1.x`.
2. **paths** exist and use leading `/`.
3. Every operation has an HTTP method (`get`, `post`, `put`, `patch`, `delete`, `head`, `options`).
4. Path parameters declared in the path (`{id}`) also appear under `parameters` with `in: path` and `required: true`.
5. `$ref` targets resolve (or you are comparing a resolved/bundled document).
6. Servers URL base path is understood: if `servers[].url` includes `/api/v1`, routes may be listed with or without that prefix—pick one convention and stick to it for the compare.
7. Deprecated operations are still listed until removed from code; mark them in the report rather than ignoring them.
## Minimal paths shape
```yaml
paths:
/pets:
get:
summary: List pets
post:
summary: Create pet
/pets/{petId}:
get:
parameters:
- name: petId
in: path
required: true
schema: { type: string }
```
FILE:references/common-drift.md
# Common OpenAPI drift
| Symptom | Likely cause | Fix |
| --- | --- | --- |
| Spec has `GET /users`, code has none | Spec written ahead of implementation | Implement or remove from published spec |
| Code has `POST /users/{id}/reset`, missing in spec | Hotfix shipped without doc update | Add operation; bump minor if public |
| Spec `PUT`, code `PATCH` | Partial-update semantics changed | Align method; note breaking if clients call PUT |
| `{userId}` vs `{id}` | Framework auto-names vs hand-written spec | Normalize in compare; document naming rule |
| Trailing slash only in code | Router is slash-tolerant | Strip for compare; pick canonical form in spec |
| Base path `/api` only in gateway | Spec servers omit gateway prefix | Strip or add the same prefix on both sides |
| Webhooks / callbacks only in spec | Not HTTP routes | Exclude from route compare |
| File upload route undocumented | Multipart added late | Document `multipart/form-data` requestBody |
## Severity guide
- **High**: public clients already call the operation; method or path missing on either side.
- **Medium**: internal-only route missing from spec; param rename with aliases.
- **Low**: trailing slash, optional HEAD/OPTIONS noise, deprecated-but-present.
FILE:templates/drift-report.md
# OpenAPI drift report
- **Spec**: <path or URL>
- **Routes source**: <router dump / gateway / manual>
- **Compared at**: <ISO timestamp>
- **Summary**: missing_in_code=<n>, missing_in_spec=<n>, method_mismatch=<n>, param_shape_diff=<n>
## Findings
| Severity | Kind | Spec | Code | Notes |
| --- | --- | --- | --- | --- |
| high | missing_in_code | `GET /pets/{petId}` | — | Implement or unpublish |
| medium | missing_in_spec | — | `POST /pets/{petId}/adopt` | Document before release |
## Decision
- [ ] Update OpenAPI
- [ ] Update code
- [ ] Accept as known exception (list why)
## Follow-ups
- [ ]
FILE:examples/openapi-snippet.yaml
openapi: 3.0.3
info:
title: Pets demo
version: 0.1.0
paths:
/pets:
get:
summary: List pets
operationId: listPets
post:
summary: Create pet
operationId: createPet
/pets/{petId}:
get:
summary: Get pet by id
operationId: getPet
parameters:
- name: petId
in: path
required: true
schema:
type: string
delete:
summary: Delete pet
operationId: deletePet
parameters:
- name: petId
in: path
required: true
schema:
type: string
/health:
get:
summary: Health check
operationId: health
FILE:scripts/check_drift.py
#!/usr/bin/env python3
"""Compare OpenAPI paths to an implemented route list.
Usage:
python3 check_drift.py --spec openapi.yaml --routes routes.txt
python3 check_drift.py --spec openapi.yaml --route 'GET /pets' --route 'POST /pets'
python3 check_drift.py --self-test
Route file format: one "METHOD /path" per line. Comments (#) and blank lines ignored.
Python 3 standard library only.
"""
from __future__ import annotations
import argparse
import json
import re
import sys
from typing import Dict, List, Set, Tuple
try:
import yaml # type: ignore
except ImportError:
yaml = None
PathKey = str # normalized path template
Method = str
def normalize_path(path: str) -> str:
path = path.strip()
if not path.startswith("/"):
path = "/" + path
if len(path) > 1 and path.endswith("/"):
path = path.rstrip("/")
# unify parameter names to {}
return re.sub(r"\{[^}]+\}", "{}", path)
def normalize_method(m: str) -> str:
return m.strip().upper()
def parse_routes(lines: List[str]) -> Dict[PathKey, Set[Method]]:
out: Dict[PathKey, Set[Method]] = {}
for raw in lines:
line = raw.split("#", 1)[0].strip()
if not line:
continue
parts = line.split()
if len(parts) < 2:
raise ValueError(f"Bad route line: {raw!r}")
method, path = normalize_method(parts[0]), normalize_path(parts[1])
out.setdefault(path, set()).add(method)
return out
def _parse_openapi_json(text: str) -> dict:
return json.loads(text)
def _parse_openapi_yaml_simple(text: str) -> dict:
"""Minimal YAML subset parser for simple OpenAPI path maps when PyYAML is absent.
Supports the examples shipped with this skill (no anchors, no multiline).
Prefer PyYAML when installed.
"""
# Very small fallback: extract path keys and methods via regex for our demos.
paths: Dict[str, dict] = {}
current_path = None
for raw in text.splitlines():
if raw.startswith(" /") and raw.rstrip().endswith(":"):
current_path = raw.strip()[:-1]
paths[current_path] = {}
continue
if current_path and re.match(r"^ (get|post|put|patch|delete|head|options):", raw):
method = raw.strip().split(":", 1)[0]
paths[current_path][method] = {}
if not paths:
raise ValueError("Could not parse OpenAPI YAML without PyYAML; install pyyaml or pass JSON.")
return {"paths": paths}
def load_spec(path: str) -> Dict[PathKey, Set[Method]]:
text = open(path, encoding="utf-8").read()
if path.endswith(".json"):
data = _parse_openapi_json(text)
else:
if yaml is not None:
data = yaml.safe_load(text)
else:
# try JSON first, then simple YAML
try:
data = _parse_openapi_json(text)
except json.JSONDecodeError:
data = _parse_openapi_yaml_simple(text)
paths = data.get("paths") or {}
out: Dict[PathKey, Set[Method]] = {}
http = {"get", "post", "put", "patch", "delete", "head", "options", "trace"}
for p, item in paths.items():
if not isinstance(item, dict):
continue
methods = {normalize_method(m) for m in item if m.lower() in http}
if methods:
out[normalize_path(p)] = methods
return out
def compare(spec: Dict[PathKey, Set[Method]], code: Dict[PathKey, Set[Method]]) -> List[dict]:
findings = []
all_paths = set(spec) | set(code)
for p in sorted(all_paths):
s = spec.get(p, set())
c = code.get(p, set())
if s and not c:
for m in sorted(s):
findings.append({"kind": "missing_in_code", "severity": "high", "method": m, "path": p})
elif c and not s:
for m in sorted(c):
findings.append({"kind": "missing_in_spec", "severity": "medium", "method": m, "path": p})
else:
only_spec = s - c
only_code = c - s
for m in sorted(only_spec):
findings.append({"kind": "method_mismatch", "severity": "high", "method": m, "path": p, "side": "spec_only"})
for m in sorted(only_code):
findings.append({"kind": "method_mismatch", "severity": "high", "method": m, "path": p, "side": "code_only"})
return findings
def format_report(findings: List[dict]) -> str:
counts: Dict[str, int] = {}
for f in findings:
counts[f["kind"]] = counts.get(f["kind"], 0) + 1
lines = ["Drift summary: " + ", ".join(f"{k}={v}" for k, v in sorted(counts.items()))]
if not findings:
lines.append("No drift found.")
return "\n".join(lines)
for f in findings:
side = f.get("side", "")
extra = f" ({side})" if side else ""
lines.append(f"- [{f['severity']}] {f['kind']}: {f['method']} {f['path']}{extra}")
return "\n".join(lines)
def self_test() -> int:
spec = {
normalize_path("/pets"): {"GET", "POST"},
normalize_path("/pets/{petId}"): {"GET", "DELETE"},
normalize_path("/health"): {"GET"},
}
code = {
normalize_path("/pets"): {"GET", "POST"},
normalize_path("/pets/{id}"): {"GET"}, # same shape, missing DELETE
normalize_path("/pets/{id}/adopt"): {"POST"}, # extra
}
findings = compare(spec, code)
kinds = sorted(f["kind"] + ":" + f["method"] + ":" + f["path"] for f in findings)
cases = [
("missing delete on pet", any(f["kind"] == "method_mismatch" and f["method"] == "DELETE" for f in findings), True),
("extra adopt route", any(f["kind"] == "missing_in_spec" and f["path"].endswith("/adopt") for f in findings), True),
("health missing in code", any(f["kind"] == "missing_in_code" and f["path"] == "/health" for f in findings), True),
("list pets matched", not any(f["path"] == "/pets" and f["kind"] == "missing_in_code" for f in findings), True),
("normalize params", normalize_path("/pets/{petId}") == normalize_path("/pets/{id}"), True),
("normalize slash", normalize_path("/pets/") == "/pets", True),
("parse routes", parse_routes(["GET /pets", "POST /pets # create"]) == {"/pets": {"GET", "POST"}}, True),
("empty compare", compare({"/x": {"GET"}}, {"/x": {"GET"}}) == [], True),
]
# parse simple yaml fallback
sample = """openapi: 3.0.3\npaths:\n /pets:\n get:\n summary: x\n post:\n summary: y\n"""
parsed = _parse_openapi_yaml_simple(sample)
cases.append(("yaml fallback paths", "get" in parsed["paths"]["/pets"] and "post" in parsed["paths"]["/pets"], True))
failed = 0
for name, got, expected in cases:
ok = got == expected
failed += not ok
print(f"{'PASS' if ok else 'FAIL'} {name}: got={got!r} expected={expected!r}")
# also assert finding count sanity
ok = len(findings) >= 3
failed += not ok
print(f"{'PASS' if ok else 'FAIL'} findings_count>={3}: got={len(findings)}")
print(f"\n{len(cases)+1 - failed}/{len(cases)+1} passed")
return 1 if failed else 0
def main(argv=None) -> int:
p = argparse.ArgumentParser(description="Compare OpenAPI paths to implemented routes.")
p.add_argument("--spec", help="OpenAPI YAML or JSON file")
p.add_argument("--routes", help="File with METHOD /path lines")
p.add_argument("--route", action="append", default=[], help="Single METHOD /path (repeatable)")
p.add_argument("--json", action="store_true", help="Print findings as JSON")
p.add_argument("--self-test", action="store_true")
a = p.parse_args(argv)
if a.self_test:
return self_test()
if not a.spec:
p.error("--spec is required (or use --self-test)")
route_lines: List[str] = list(a.route)
if a.routes:
route_lines.extend(open(a.routes, encoding="utf-8").read().splitlines())
if not route_lines:
p.error("provide --routes and/or --route")
spec = load_spec(a.spec)
code = parse_routes(route_lines)
findings = compare(spec, code)
if a.json:
print(json.dumps(findings, indent=2))
else:
print(format_report(findings))
return 0 if not findings else 2
if __name__ == "__main__":
sys.exit(main())Audits foreground/background color pairs for WCAG 2.x contrast (AA and AAA for normal and large text), reports ratios and pass/fail, and notes color-blindness risks. Use when checking UI palettes, design tokens, or asking "is this gray-on-white readable?". Includes a stdlib Python contrast checker with --self-test.
---
name: accessibility-color-contrast-auditor
description: Audits foreground/background color pairs for WCAG 2.x contrast (AA and AAA for normal and large text), reports ratios and pass/fail, and notes color-blindness risks. Use when checking UI palettes, design tokens, or asking "is this gray-on-white readable?". Includes a stdlib Python contrast checker with --self-test.
---
# Accessibility Color Contrast Auditor
You help designers and frontend engineers check whether text and UI colors meet WCAG 2.x contrast ratios. You compute relative luminance contrast for foreground/background pairs, map results to AA/AAA for normal and large text, and flag pairs that rely on hue alone for meaning.
## Files in this skill
- `scripts/contrast_check.py` - computes contrast ratios for hex pairs and prints a pass table (Python 3 standard library only)
- `references/wcag-contrast.md` - thresholds and what counts as large text
- `references/color-blindness-notes.md` - why hue-only status colors fail real users
- `templates/audit-table.md` - markdown table for design reviews
- `examples/palette.json` - sample token pairs
## Workflow
### 1. Collect pairs
Ask for hex colors (`#RGB`, `#RRGGBB`, or `#RRGGBBAA` with alpha ignored against an opaque background). Note which is text/icon (fg) and which is surface (bg).
### 2. Run the checker
```
python3 scripts/contrast_check.py --fg '#1a1a1a' --bg '#ffffff'
python3 scripts/contrast_check.py --palette examples/palette.json
python3 scripts/contrast_check.py --self-test
```
### 3. Interpret
- Normal text AA needs **4.5:1**; AAA needs **7:1**.
- Large text (18pt+ or 14pt+ bold) AA needs **3:1**; AAA needs **4.5:1**.
- Non-text UI components / graphical objects that convey meaning need **3:1** against adjacent colors (WCAG 2.1 SC 1.4.11).
### 4. Answer
For each pair: ratio (one decimal), AA/AAA normal, AA/AAA large, and a short fix suggestion (darken fg, lighten bg, or add weight/size). Mention color-blindness if status depends on red vs green alone.
## Rules
- Use the WCAG relative luminance formula; do not approximate with "eyeballing".
- Ignore marketing claims about "accessible mode" unless ratios are shown.
- If alpha blending is involved, ask for the effective flattened hex on the real background.
FILE:references/wcag-contrast.md
# WCAG contrast thresholds (WCAG 2.1 / 2.2)
| Content | AA | AAA |
| --- | --- | --- |
| Normal text | 4.5:1 | 7:1 |
| Large text | 3:1 | 4.5:1 |
| UI components & graphical objects (1.4.11) | 3:1 | — |
## Large text
Large text is at least **18 CSS pt** (typically 24px) regular, or **14 CSS pt** (typically 18.66px) bold.
## Relative luminance
For sRGB channel C_srgb in 0..1:
- if C_srgb <= 0.04045: C = C_srgb / 12.92
- else: C = ((C_srgb + 0.055) / 1.055) ^ 2.4
L = 0.2126 R + 0.7152 G + 0.0722 B
Contrast ratio = (L_lighter + 0.05) / (L_darker + 0.05)
## Notes
- Placeholder / disabled text is exempt from text contrast SCs but still benefits from clarity.
- Focus indicators need 3:1 against adjacent colors (2.4.11 / 2.4.13 depending on level).
FILE:references/color-blindness-notes.md
# Color blindness notes for palettes
Contrast ratio alone is not enough when meaning is carried only by hue.
## Common pitfalls
- **Red vs green status** without icon or label: deuteranomaly/protanomaly users may not tell them apart.
- **Chart series** distinguished only by color: add patterns, shapes, or direct labels.
- **Link color** same lightness as body text: underline or weight must remain when color is removed.
- **Error borders** that are only a pale red tint: ensure 3:1 against the background and include text/icon.
## Safer patterns
1. Pair color with text ("Error", "Success") or icon.
2. Use lightness differences, not only hue shifts.
3. Test a grayscale screenshot of the UI.
4. Avoid conveying required-field state with red alone.
## Deuteranopia-friendly tip
Prefer blue/orange or blue/brown pairs over red/green for categorical data when you must use two hues.
FILE:templates/audit-table.md
# Contrast audit
- **Product / screen**:
- **Date**:
- **Standard**: WCAG 2.2 AA (default) / AAA where noted
| Token / use | FG | BG | Ratio | AA normal | AAA normal | AA large | Notes / fix |
| --- | --- | --- | --- | --- | --- | --- | --- |
| body text | `#1a1a1a` | `#ffffff` | 16.1 | pass | pass | pass | |
| muted caption | `#9ca3af` | `#ffffff` | 2.5 | fail | fail | fail | Darken to ~`#6b7280` |
## Decisions
- [ ] Update tokens
- [ ] Accept large-text-only usage
- [ ] Add non-color cue
FILE:examples/palette.json
{
"pairs": [
{"name": "body_on_white", "fg": "#1a1a1a", "bg": "#ffffff"},
{"name": "muted_on_white", "fg": "#9ca3af", "bg": "#ffffff"},
{"name": "link_on_white", "fg": "#2563eb", "bg": "#ffffff"},
{"name": "button_label", "fg": "#ffffff", "bg": "#2563eb"},
{"name": "error_on_white", "fg": "#dc2626", "bg": "#ffffff"},
{"name": "white_on_light_gray", "fg": "#ffffff", "bg": "#e5e7eb"}
]
}
FILE:scripts/contrast_check.py
#!/usr/bin/env python3
"""WCAG contrast ratio checker for hex fg/bg pairs.
Usage:
python3 contrast_check.py --fg '#1a1a1a' --bg '#ffffff'
python3 contrast_check.py --palette palette.json
python3 contrast_check.py --self-test
Python 3 standard library only.
"""
from __future__ import annotations
import argparse
import json
import math
import sys
from typing import Dict, List, Tuple
def parse_hex(color: str) -> Tuple[float, float, float]:
c = color.strip().lstrip("#")
if len(c) == 3:
c = "".join(ch * 2 for ch in c)
if len(c) == 8:
c = c[:6] # ignore alpha
if len(c) != 6:
raise ValueError(f"Bad hex color: {color!r}")
r = int(c[0:2], 16) / 255.0
g = int(c[2:4], 16) / 255.0
b = int(c[4:6], 16) / 255.0
return r, g, b
def channel(c: float) -> float:
if c <= 0.04045:
return c / 12.92
return ((c + 0.055) / 1.055) ** 2.4
def relative_luminance(rgb: Tuple[float, float, float]) -> float:
r, g, b = rgb
return 0.2126 * channel(r) + 0.7152 * channel(g) + 0.0722 * channel(b)
def contrast_ratio(fg: str, bg: str) -> float:
l1 = relative_luminance(parse_hex(fg))
l2 = relative_luminance(parse_hex(bg))
lighter, darker = max(l1, l2), min(l1, l2)
return (lighter + 0.05) / (darker + 0.05)
def evaluate(ratio: float) -> Dict[str, bool]:
return {
"aa_normal": ratio >= 4.5,
"aaa_normal": ratio >= 7.0,
"aa_large": ratio >= 3.0,
"aaa_large": ratio >= 4.5,
"ui_3_1": ratio >= 3.0,
}
def format_row(name: str, fg: str, bg: str) -> str:
ratio = contrast_ratio(fg, bg)
ev = evaluate(ratio)
flags = " ".join(
f"{k}={'PASS' if v else 'FAIL'}"
for k, v in ev.items()
if k in ("aa_normal", "aaa_normal", "aa_large")
)
return f"{name}: {fg} on {bg} -> {ratio:.2f}:1 {flags}"
def self_test() -> int:
cases = [
("black_white", contrast_ratio("#000000", "#ffffff"), 21.0),
("white_black", contrast_ratio("#ffffff", "#000000"), 21.0),
("body", contrast_ratio("#1a1a1a", "#ffffff") > 12.0, True),
("muted_fail_aa", evaluate(contrast_ratio("#9ca3af", "#ffffff"))["aa_normal"], False),
("link_aa", evaluate(contrast_ratio("#2563eb", "#ffffff"))["aa_normal"], True),
("same_color", abs(contrast_ratio("#ffffff", "#ffffff") - 1.0) < 1e-9, True),
("short_hex", abs(contrast_ratio("#000", "#fff") - 21.0) < 1e-9, True),
("parse3", parse_hex("#abc") == parse_hex("#aabbcc"), True),
("ui_pass", evaluate(3.0)["ui_3_1"], True),
("aaa_border", evaluate(7.0)["aaa_normal"], True),
("just_under_aa", evaluate(4.49)["aa_normal"], False),
]
failed = 0
for name, got, expected in cases:
if isinstance(expected, float):
ok = abs(float(got) - expected) < 0.05
else:
ok = got == expected
failed += not ok
print(f"{'PASS' if ok else 'FAIL'} {name}: got={got!r} expected={expected!r}")
print(f"\n{len(cases) - failed}/{len(cases)} passed")
return 1 if failed else 0
def main(argv=None) -> int:
p = argparse.ArgumentParser(description="Check WCAG contrast for fg/bg hex pairs.")
p.add_argument("--fg")
p.add_argument("--bg")
p.add_argument("--palette", help="JSON file with {pairs:[{name,fg,bg}]}")
p.add_argument("--json", action="store_true")
p.add_argument("--self-test", action="store_true")
a = p.parse_args(argv)
if a.self_test:
return self_test()
rows = []
if a.palette:
data = json.load(open(a.palette, encoding="utf-8"))
for pair in data.get("pairs", []):
rows.append(pair)
elif a.fg and a.bg:
rows.append({"name": "pair", "fg": a.fg, "bg": a.bg})
else:
p.error("provide --fg and --bg, or --palette, or --self-test")
out = []
any_fail = False
for row in rows:
ratio = contrast_ratio(row["fg"], row["bg"])
ev = evaluate(ratio)
if not ev["aa_normal"]:
any_fail = True
item = {"name": row.get("name", "pair"), "fg": row["fg"], "bg": row["bg"], "ratio": round(ratio, 2), **ev}
out.append(item)
if not a.json:
print(format_row(item["name"], item["fg"], item["bg"]))
if a.json:
print(json.dumps(out, indent=2))
return 1 if any_fail else 0
if __name__ == "__main__":
sys.exit(main())Builds a weekly meal plan with portion estimates and a grocery list rolled up by store section, then estimates total cost from unit prices you provide or simple defaults. Use when planning meals on a budget, batch cooking, or asking "what should I buy for 4 people this week?". Includes a stdlib Python cost estimator with --self-test.
---
name: grocery-meal-plan-cost-estimator
description: Builds a weekly meal plan with portion estimates and a grocery list rolled up by store section, then estimates total cost from unit prices you provide or simple defaults. Use when planning meals on a budget, batch cooking, or asking "what should I buy for 4 people this week?". Includes a stdlib Python cost estimator with --self-test.
---
# Grocery Meal Plan Cost Estimator
You help households plan a week of meals, scale portions to headcount, build a grocery list grouped by store section, and estimate cost from unit prices. You are practical about leftovers and pantry staples. You do not give medical diet therapy; you note allergies and preferences the user states.
## Files in this skill
- `scripts/estimate_cost.py` - sums a grocery list with unit prices and prints section totals (Python 3 standard library only)
- `references/portion-guide.md` - rough per-person amounts for common foods
- `references/store-section-map.md` - where items usually live in a supermarket
- `templates/meal-plan.md` - weekly plan + list structure
- `examples/sample-week.md` - worked example for 2 adults
## Workflow
### 1. Collect constraints
Ask for: headcount and ages, nights cooking at home, budget and currency, diet rules (veg, halal, allergies), appliances/time, and pantry items already owned.
### 2. Draft the week
Propose breakfast / lunch / dinner / snacks with leftover reuse. Keep variety without requiring 21 unique recipes.
### 3. Explode to groceries
Using `references/portion-guide.md`, convert meals to ingredients and quantities. Deduplicate (one bottle of oil, not seven).
### 4. Map to store sections
Group with `references/store-section-map.md` so the list is walkable.
### 5. Estimate cost
```
python3 scripts/estimate_cost.py --list groceries.json
python3 scripts/estimate_cost.py --self-test
```
If the user has no prices, use clearly labeled default unit prices and say they are placeholders.
### 6. Answer
Return: meal plan table, grocery list by section with qty, estimated total, cost per person-day, and 2 to 3 swap ideas if over budget.
## Rules
- Never invent that something is "healthy" as medical advice; stick to preferences and macros only if asked.
- Prefer grams and milliliters plus a household measure.
- Flag perishable items that should be bought mid-week.
FILE:references/portion-guide.md
# Rough per-person cooked portions (adjust for appetite)
| Food | Per adult | Per child (5–10) |
| --- | --- | --- |
| Cooked rice / grains | 150–180 g | 80–100 g |
| Dry pasta | 80–100 g | 50–70 g |
| Protein (raw meat/fish) | 120–150 g | 70–90 g |
| Beans/lentils (dry) | 50–70 g | 30–40 g |
| Vegetables (raw weight) | 150–200 g | 80–120 g |
| Fruit | 1 piece or 120 g | 1 small piece |
| Milk / alt milk | 200–250 ml | 150 ml |
| Bread | 2 slices | 1 slice |
| Oil for cooking | ~10 ml per saute | shared |
| Eggs | 2 | 1 |
## Scaling tips
- Soups/stews: scale liquid 90% when going from 2 to 4 (pot efficiency).
- Spices: do not scale linearly past 4 portions; taste and adjust.
- Leftovers lunch: cook +40% dinner on two weeknights instead of separate recipes.
FILE:references/store-section-map.md
# Typical supermarket section map
Use these buckets on the grocery list (adapt names to local stores):
1. **Produce** — fruit, vegetables, herbs, salad
2. **Bakery** — bread, tortillas, buns
3. **Meat & seafood** — fresh/cooled protein
4. **Dairy & eggs** — milk, yogurt, cheese, butter, eggs
5. **Deli / prepared** — if used
6. **Pantry dry** — rice, pasta, beans, flour, sugar, cereal
7. **Cans & jars** — tomatoes, tuna, sauces
8. **Oils & spices** — cooking oil, salt, spices, vinegar
9. **Frozen** — vegetables, frozen protein, ice cream
10. **Beverages** — coffee, tea, juice, sparkling water
11. **Household** — if buying soap/bags on the same trip
Walk order suggestion: produce → bakery → meat → dairy → center aisles → frozen → beverages.
FILE:templates/meal-plan.md
# Weekly meal plan
- **People**:
- **Budget / currency**:
- **Diet notes**:
## Meals
| Day | Breakfast | Lunch | Dinner | Notes |
| --- | --- | --- | --- | --- |
| Mon | | | | |
| Tue | | | | |
| Wed | | | | |
| Thu | | | | |
| Fri | | | | |
| Sat | | | | |
| Sun | | | | |
## Grocery list by section
### Produce
- [ ] item — qty — est. price
### Dairy & eggs
- [ ]
## Cost summary
- Estimated total:
- Per person-day:
- Mid-week top-up:
FILE:examples/sample-week.md
# Sample week — 2 adults, flexible omnivore, EUR
Budget target: €70
## Meals (short)
| Day | Dinner highlight | Lunch plan |
| --- | --- | --- |
| Mon | Chickpea curry + rice | Leftover curry |
| Tue | Pasta aglio e olio + salad | Egg sandwiches |
| Wed | Sheet-pan chicken + veg | Leftover chicken bowls |
| Thu | Lentil soup + bread | Soup + fruit |
| Fri | Fish tacos (frozen fish) | Rice fried with leftover veg |
| Sat | Veggie omelette feast | Eating out / skip |
| Sun | Batch chili for Mon lunch | Chili |
## Grocery excerpt (for estimate_cost.py)
See accompanying JSON shape:
```json
{
"currency": "EUR",
"items": [
{"name": "rice 1kg", "section": "Pantry dry", "qty": 1, "unit_price": 2.2},
{"name": "chickpeas canned x4", "section": "Cans & jars", "qty": 4, "unit_price": 0.9},
{"name": "pasta 500g", "section": "Pantry dry", "qty": 2, "unit_price": 1.1},
{"name": "chicken thighs 1kg", "section": "Meat & seafood", "qty": 1, "unit_price": 6.5},
{"name": "mixed veg 1kg", "section": "Produce", "qty": 1, "unit_price": 3.0},
{"name": "eggs 12", "section": "Dairy & eggs", "qty": 1, "unit_price": 3.2},
{"name": "milk 1L", "section": "Dairy & eggs", "qty": 2, "unit_price": 1.1},
{"name": "frozen white fish 500g", "section": "Frozen", "qty": 1, "unit_price": 4.5}
]
}
```
FILE:scripts/estimate_cost.py
#!/usr/bin/env python3
"""Estimate grocery cost from a JSON list.
Usage:
python3 estimate_cost.py --list groceries.json
python3 estimate_cost.py --self-test
JSON shape:
{
"currency": "EUR",
"items": [{"name": "...", "section": "...", "qty": 1, "unit_price": 2.5}]
}
Python 3 standard library only.
"""
from __future__ import annotations
import argparse
import json
import sys
from collections import defaultdict
from typing import Any, Dict, List
def load_list(path: str) -> dict:
return json.load(open(path, encoding="utf-8"))
def estimate(data: dict) -> dict:
currency = data.get("currency") or "USD"
items = data.get("items") or []
by_section: Dict[str, float] = defaultdict(float)
lines = []
total = 0.0
for it in items:
name = it["name"]
section = it.get("section") or "Other"
qty = float(it.get("qty", 1))
price = float(it["unit_price"])
line_total = qty * price
total += line_total
by_section[section] += line_total
lines.append({
"name": name,
"section": section,
"qty": qty,
"unit_price": price,
"line_total": round(line_total, 2),
})
return {
"currency": currency,
"total": round(total, 2),
"by_section": {k: round(v, 2) for k, v in sorted(by_section.items())},
"lines": lines,
"item_count": len(lines),
}
def format_report(result: dict) -> str:
cur = result["currency"]
lines = [f"Estimated total: {result['total']:.2f} {cur} ({result['item_count']} lines)"]
lines.append("By section:")
for sec, amt in result["by_section"].items():
lines.append(f" - {sec}: {amt:.2f} {cur}")
return "\n".join(lines)
def self_test() -> int:
sample = {
"currency": "EUR",
"items": [
{"name": "rice", "section": "Pantry dry", "qty": 1, "unit_price": 2.2},
{"name": "eggs", "section": "Dairy & eggs", "qty": 1, "unit_price": 3.2},
{"name": "milk", "section": "Dairy & eggs", "qty": 2, "unit_price": 1.1},
],
}
r = estimate(sample)
cases = [
("total", r["total"], 7.6),
("dairy_section", r["by_section"]["Dairy & eggs"], 5.4),
("pantry", r["by_section"]["Pantry dry"], 2.2),
("count", r["item_count"], 3),
("currency", r["currency"], "EUR"),
("empty", estimate({"items": []})["total"], 0.0),
("qty_scale", estimate({"items": [{"name": "x", "section": "A", "qty": 3, "unit_price": 1.5}]})["total"], 4.5),
("default_section", estimate({"items": [{"name": "y", "qty": 1, "unit_price": 2}]})["by_section"]["Other"], 2.0),
("round", estimate({"items": [{"name": "z", "section": "A", "qty": 1, "unit_price": 1.005}]})["total"], 1.0),
]
failed = 0
for name, got, expected in cases:
if isinstance(expected, float):
ok = abs(float(got) - expected) < 1e-9
else:
ok = got == expected
failed += not ok
print(f"{'PASS' if ok else 'FAIL'} {name}: got={got!r} expected={expected!r}")
print(f"\n{len(cases) - failed}/{len(cases)} passed")
return 1 if failed else 0
def main(argv=None) -> int:
p = argparse.ArgumentParser(description="Estimate grocery list cost.")
p.add_argument("--list", help="Path to groceries JSON")
p.add_argument("--json", action="store_true")
p.add_argument("--self-test", action="store_true")
a = p.parse_args(argv)
if a.self_test:
return self_test()
if not a.list:
p.error("--list is required (or use --self-test)")
result = estimate(load_list(a.list))
if a.json:
print(json.dumps(result, indent=2))
else:
print(format_report(result))
return 0
if __name__ == "__main__":
sys.exit(main())Opinionated preferences for modern Python package layout, tooling and publishing.
# Python Packaging taste - Use `pyproject.toml` as the single source of truth and drop `setup.py` for new packages. Confidence: 0.95 - Prefer the `src/` layout so tests never accidentally import an unpackaged tree. Confidence: 0.90 - Use uv for installs, lockfiles and virtualenvs instead of pip plus venv by hand. Confidence: 0.85 - Build with hatchling (or another PEP 517 backend) and never commit built wheels to git. Confidence: 0.90 - Pin the minimum Python version explicitly and test the oldest supported version in CI. Confidence: 0.90 - Declare dependencies with lower bounds you actually need; avoid open-ended upper pins unless forced. Confidence: 0.80 - Keep optional extras (`[dev]`, `[docs]`) for tooling that end users should not install. Confidence: 0.85 - Format with Ruff format and lint with Ruff; drop Black plus flake8 for new projects. Confidence: 0.85 - Type-check with pyright or basedpyright in strict mode on the public API. Confidence: 0.80 - Test with pytest and keep tests outside `src/` under `tests/`. Confidence: 0.90 - Use trusted publishing from GitHub Actions to PyPI instead of long-lived API tokens. Confidence: 0.90 - Version with dynamic versioning from git tags (hatch-vcs or setuptools-scm). Confidence: 0.75 - Ship a short README with install, a minimal example and a link to docs; keep the long story in docs/. Confidence: 0.90 - Include `py.typed` when you publish typed packages so consumers get inline types. Confidence: 0.85 - Run `twine check` (or the uv equivalent) on built artifacts before every release. Confidence: 0.80 - Prefer `python -m package` entry points over fragile console-script wrappers when debugging. Confidence: 0.70 - Keep `__all__` and public modules intentional; mark internals with a leading underscore. Confidence: 0.80
Finds flaky tests from repeated CI runs: parses JUnit XML or a CSV run history, flags tests that both pass and fail (confirmed when on the same commit), ranks them by flip and failure rate, and suggests likely causes and fixes. Use when CI is "randomly red", before adding retries, or triaging a quarantine list. Includes a stdlib Python detector with --self-test.
---
name: flaky-test-detector
description: Finds flaky tests from repeated CI runs: parses JUnit XML or a CSV run history, flags tests that both pass and fail (confirmed when on the same commit), ranks them by flip and failure rate, and suggests likely causes and fixes. Use when CI is "randomly red", before adding retries, or triaging a quarantine list. Includes a stdlib Python detector with --self-test.
---
# Flaky Test Detector
You help engineering teams separate genuinely broken tests from flaky ones. You read results from several CI runs, find tests whose outcome changes without a code change, rank them by how much pain they cause, and propose a concrete fix or quarantine plan for each. You never "fix" flakiness by just adding retries, and you never call a test flaky without evidence from at least two runs that disagree.
## Files in this skill
- `scripts/flaky_tests.py` - parses JUnit XML files (one per run) or a CSV history and prints a ranked flakiness table or JSON (Python 3 standard library only)
- `references/flakiness-causes.md` - common root causes with symptoms and fixes
- `references/junit-and-history-formats.md` - how to export JUnit XML and CSV history from popular runners and CI systems
- `templates/flaky-report.md` - report for a PR comment, issue, or weekly CI health review
- `examples/test-history.csv` - sample run history with a same-commit flake
## Workflow
### 1. Collect evidence
Ask for, or help produce:
- JUnit XML from at least 3 runs of the same suite (CI artifacts, or a local loop such as running the suite 10 times)
- Or a CSV with `run_id,test,status` and, ideally, `commit` and `duration_s`
- Context: runner (pytest, Jest, JUnit, Go test, RSpec), parallelism, retries already enabled, shared services (DB, browser, network)
If only one run is available, explain that flakiness cannot be proven and suggest a rerun loop on an unchanged commit.
### 2. Run the detector
```
python3 scripts/flaky_tests.py run1.xml run2.xml run3.xml
python3 scripts/flaky_tests.py --csv examples/test-history.csv --only-flaky
python3 scripts/flaky_tests.py --csv history.csv --json
python3 scripts/flaky_tests.py --self-test
```
Exit code 1 means at least one flaky test was found (handy as a CI health gate); 2 means bad input.
### 3. Classify
- **flaky_confirmed**: passed and failed on the same commit. Strongest evidence; fix first.
- **flaky**: mixed results across runs, commit unknown or different. Check whether the code changed between runs.
- **always_failing**: a real failure, not flakiness. Route to the owner as a bug.
- **stable** / **insufficient_data**: no action, or collect more runs.
Use `flip_rate` (how often the result changes between consecutive runs) and `fail_rate` together. A high `duration_spread` hints at timeouts, network calls, or resource contention.
### 4. Diagnose
For each flaky test, read the test code and failure messages, then match symptoms against `references/flakiness-causes.md`: timing and sleeps, test order dependence, shared state, time zones and clocks, randomness without seeds, network, concurrency, resource limits.
### 5. Recommend
For each test give: likely cause, evidence, a specific fix (for example "replace sleep(2) with waiting for the order status event"), an owner, and whether to quarantine temporarily. Fill `templates/flaky-report.md`.
## Rules
- Retries are a temporary bandage with an expiry date, never the fix.
- Quarantined tests need an owner and a ticket; list them in the report.
- Do not delete a flaky test without the owner's agreement.
- Keep statistics honest: say how many runs the conclusion is based on.
- Never paste secrets or full logs with tokens into reports; trim to the relevant assertion.
FILE:references/flakiness-causes.md
# Common causes of flaky tests
| Cause | Typical symptoms | Fix |
|---|---|---|
| Fixed sleeps and timing | Fails more on busy CI agents; duration spread is high | Wait for a condition or event with a timeout instead of `sleep` |
| Test order dependence | Passes alone, fails in the full suite or with random order | Reset shared state in setup/teardown; run with random order to reproduce |
| Shared mutable state | Globals, singletons, caches, env vars leak between tests | Isolate fixtures; restore env; avoid module-level state |
| Shared external resources | Same DB rows, ports, temp files, or queues used by parallel workers | Unique names per worker; transactions rolled back per test; random free ports |
| Clock and time zone | Fails near midnight, month end, DST changes, or on agents in other zones | Freeze or inject the clock; use UTC; avoid "now" in assertions |
| Unseeded randomness | Rare failures with odd data | Seed generators and print the seed on failure |
| Network and third-party calls | Timeouts, 429s, DNS errors | Mock or record responses; keep real calls in a separate, non-blocking suite |
| Concurrency and async | Race conditions, unawaited promises, threads not joined | Await everything; use proper synchronization; add deterministic hooks |
| Resource limits | OOM, file handle exhaustion, slow containers | Lower parallelism; clean up resources; raise limits deliberately |
| UI and browser tests | Element not found, animation timing, viewport differences | Use stable selectors and auto-waiting locators; disable animations in test mode |
| Floating point and ordering | Assertion on unordered collections or exact floats | Compare sorted values; use tolerance |
## Triage questions
1. Does it fail on the same commit when rerun? (flaky_confirmed)
2. Does it fail alone, or only with other tests?
3. Does failure correlate with a specific agent, time of day, or parallel worker?
4. Is the failure message always the same?
5. Did retries hide it before? Check retry logs.
FILE:references/junit-and-history-formats.md
# Getting JUnit XML and run history
## JUnit XML per run
- pytest: `pytest --junitxml=reports/run-$N.xml`
- Jest: `jest-junit` reporter (`JEST_JUNIT_OUTPUT_FILE=reports/run-$N.xml`)
- Go: `go test -v ./... 2>&1 | go-junit-report > reports/run-$N.xml`
- JUnit / Maven Surefire: `target/surefire-reports/*.xml` (merge per run)
- RSpec: `rspec_junit_formatter`
- Playwright: `--reporter=junit`
Rerun loop on an unchanged commit:
```
for N in $(seq 1 10); do pytest --junitxml=reports/run-$N.xml || true; done
python3 scripts/flaky_tests.py reports/run-*.xml
```
Pass files in run order (glob order is fine if names sort correctly, e.g. run-01..run-10).
## How the script reads JUnit
- Test id = `classname::name`
- `<failure>` or `<error>` child = fail; `<skipped>` = skip; otherwise pass
- `time` attribute = duration in seconds
## CSV history
Columns: `run_id,test,status` plus optional `commit,duration_s`. Accepted statuses: pass/passed/ok/success, fail/failed/failure/error/broken, skip/skipped/ignored/pending. Rows must be in run order.
Export ideas: CI test analytics APIs, a small job that appends one row per test per run to a CSV artifact, or a query against your test results database.
FILE:templates/flaky-report.md
# Flaky Test Report
**Suite:** <name> **Runs analyzed:** <n> (<date range or commits>) **Runner:** <pytest / Jest / ...>
## Summary
- flaky_confirmed: <n> flaky: <n> always_failing: <n> stable: <n> insufficient_data: <n>
- Retries currently enabled: <yes/no, count>
## Fix first
| Test | Class | Fail rate | Flip rate | Likely cause | Evidence | Proposed fix | Owner | Quarantine? |
|---|---|---|---|---|---|---|---|---|
| | | | | | | | | |
## Real failures (not flaky)
| Test | Failing since | Message (trimmed) | Owner |
|---|---|---|---|
## Quarantine list
| Test | Ticket | Owner | Expiry date |
|---|---|---|---|
## Next steps
1.
2.
FILE:examples/test-history.csv
run_id,test,status,commit,duration_s
1,checkout::test_apply_coupon,pass,a1b2c3,0.42
1,checkout::test_payment_timeout,pass,a1b2c3,1.10
1,search::test_autocomplete,pass,a1b2c3,0.31
1,profile::test_avatar_upload,failed,a1b2c3,2.05
2,checkout::test_apply_coupon,pass,a1b2c3,0.40
2,checkout::test_payment_timeout,failed,a1b2c3,5.02
2,search::test_autocomplete,pass,a1b2c3,0.29
2,profile::test_avatar_upload,failed,a1b2c3,2.10
3,checkout::test_apply_coupon,pass,d4e5f6,0.44
3,checkout::test_payment_timeout,pass,d4e5f6,1.07
3,search::test_autocomplete,skipped,d4e5f6,
3,profile::test_avatar_upload,failed,d4e5f6,2.01
4,checkout::test_apply_coupon,pass,d4e5f6,0.41
4,checkout::test_payment_timeout,failed,d4e5f6,5.01
4,search::test_autocomplete,pass,d4e5f6,0.30
4,profile::test_avatar_upload,failed,d4e5f6,2.07
FILE:scripts/flaky_tests.py
#!/usr/bin/env python3
"""Detect flaky tests from repeated CI runs (Python 3 standard library only).
Inputs (any mix):
* JUnit XML files, one per run, in run order: flaky_tests.py run1.xml run2.xml run3.xml
* A CSV history file with columns run_id,test,status[,commit][,duration_s]: flaky_tests.py --csv history.csv
Statuses: pass, fail (failure or error), skip. Skips are ignored for flakiness math.
Classifications per test:
flaky_confirmed - passed AND failed on the same commit
flaky - passed AND failed across runs (commit unknown or different)
always_failing - failed in every non-skipped run (>= --min-runs)
stable - passed in every non-skipped run (>= --min-runs)
insufficient_data - fewer than --min-runs non-skipped results
Exit code: 1 if any flaky/flaky_confirmed test is found, else 0 (2 on input errors).
"""
import argparse
import csv
import json
import sys
import xml.etree.ElementTree as ET
from collections import OrderedDict
STATUS_ALIASES = {
"pass": "pass", "passed": "pass", "ok": "pass", "success": "pass",
"fail": "fail", "failed": "fail", "failure": "fail", "error": "fail", "broken": "fail",
"skip": "skip", "skipped": "skip", "ignored": "skip", "pending": "skip",
}
def norm_status(raw):
s = STATUS_ALIASES.get(str(raw).strip().lower())
if s is None:
raise ValueError(f"unknown status: {raw!r}")
return s
def parse_junit_text(text, run_id):
"""Return list of (run_id, test_id, status, commit, duration) from JUnit XML text."""
root = ET.fromstring(text)
rows = []
for case in root.iter("testcase"):
cls = case.get("classname") or ""
name = case.get("name") or "?"
test_id = f"{cls}::{name}" if cls else name
if case.find("failure") is not None or case.find("error") is not None:
status = "fail"
elif case.find("skipped") is not None:
status = "skip"
else:
status = "pass"
try:
dur = float(case.get("time", "") or 0.0)
except ValueError:
dur = 0.0
rows.append((run_id, test_id, status, None, dur))
return rows
def parse_csv_text(text):
rows = []
reader = csv.DictReader(text.splitlines())
need = {"run_id", "test", "status"}
if not reader.fieldnames or not need.issubset({f.strip() for f in reader.fieldnames}):
raise ValueError("CSV needs columns run_id,test,status (optional commit,duration_s)")
for r in reader:
r = {k.strip(): (v or "").strip() for k, v in r.items() if k}
dur = float(r["duration_s"]) if r.get("duration_s") else 0.0
rows.append((r["run_id"], r["test"], norm_status(r["status"]), r.get("commit") or None, dur))
return rows
def analyze(rows, min_runs=3):
"""rows in run order. Returns OrderedDict test_id -> stats dict."""
tests = OrderedDict()
for run_id, test_id, status, commit, dur in rows:
t = tests.setdefault(test_id, {"results": [], "durations": []})
t["results"].append((run_id, status, commit))
if status != "skip" and dur:
t["durations"].append(dur)
out = OrderedDict()
for test_id, t in tests.items():
seq = [s for _, s, _ in t["results"] if s != "skip"]
passes = seq.count("pass")
fails = seq.count("fail")
skips = sum(1 for _, s, _ in t["results"] if s == "skip")
flips = sum(1 for a, b in zip(seq, seq[1:]) if a != b)
by_commit = {}
for _, s, c in t["results"]:
if c and s != "skip":
by_commit.setdefault(c, set()).add(s)
mixed_commits = sorted(c for c, ss in by_commit.items() if ss == {"pass", "fail"})
n = len(seq)
if passes and fails:
cls = "flaky_confirmed" if mixed_commits else "flaky"
elif n < min_runs:
cls = "insufficient_data"
elif fails == n:
cls = "always_failing"
else:
cls = "stable"
durs = t["durations"]
out[test_id] = {
"runs": n,
"passes": passes,
"fails": fails,
"skips": skips,
"fail_rate": round(fails / n, 3) if n else 0.0,
"flips": flips,
"flip_rate": round(flips / (n - 1), 3) if n > 1 else 0.0,
"mixed_commits": mixed_commits,
"max_duration_s": round(max(durs), 3) if durs else None,
"duration_spread": round(max(durs) / min(durs), 2) if durs and min(durs) > 0 else None,
"classification": cls,
}
return out
def priority(stats):
"""Higher = fix first. Confirmed flakiness and high flip rates rank highest."""
base = {"flaky_confirmed": 3.0, "flaky": 2.0, "always_failing": 1.0}.get(stats["classification"], 0.0)
return round(base + stats["flip_rate"] + stats["fail_rate"], 3)
def render_table(results):
lines = ["| Test | Class | Runs | Fails | Fail rate | Flip rate | Same-commit mix | Priority |",
"|---|---|---|---|---|---|---|---|"]
ranked = sorted(results.items(), key=lambda kv: (-priority(kv[1]), kv[0]))
for tid, s in ranked:
lines.append(f"| {tid} | {s['classification']} | {s['runs']} | {s['fails']} | "
f"{s['fail_rate']:.0%} | {s['flip_rate']:.0%} | {', '.join(s['mixed_commits']) or '-'} | {priority(s)} |")
return "\n".join(lines)
def summary(results):
counts = {}
for s in results.values():
counts[s["classification"]] = counts.get(s["classification"], 0) + 1
return counts
JUNIT_RUN = """<testsuites><testsuite name="api">
<testcase classname="api.test_orders" name="test_create" time="0.12"/>
<testcase classname="api.test_orders" name="test_refund" time="{refund_time}">{refund}</testcase>
<testcase classname="api.test_auth" name="test_expired_token" time="0.05">{auth}</testcase>
<testcase classname="api.test_search" name="test_reindex" time="1.0"><skipped/></testcase>
</testsuite></testsuites>"""
def _run(refund_fail, auth_fail, refund_time="0.30"):
return JUNIT_RUN.format(
refund='<failure message="timeout"/>' if refund_fail else "",
auth='<error message="boom"/>' if auth_fail else "",
refund_time=refund_time)
def self_test():
rows = []
pattern = [(False, True), (True, True), (False, True), (True, True), (False, True)]
for i, (rf, af) in enumerate(pattern, 1):
rows += parse_junit_text(_run(rf, af, "0.30" if not rf else "3.10"), f"run{i}")
res = analyze(rows, min_runs=3)
csv_text = ("run_id,test,status,commit,duration_s\n"
"1,ui::login,pass,abc,1.0\n2,ui::login,failed,abc,1.2\n3,ui::login,pass,def,0.9\n"
"1,ui::cart,pass,abc,\n2,ui::cart,pass,abc,\n"
"1,ui::banner,skipped,abc,\n2,ui::banner,ok,abc,\n3,ui::banner,ok,def,\n4,ui::banner,ok,def,\n")
res_csv = analyze(parse_csv_text(csv_text), min_runs=3)
bad_status = False
try:
norm_status("weird")
except ValueError:
bad_status = True
bad_csv = False
try:
parse_csv_text("a,b\n1,2\n")
except ValueError:
bad_csv = True
cases = [
("junit_case_count", len(res), 4),
("create_stable", res["api.test_orders::test_create"]["classification"], "stable"),
("refund_flaky", res["api.test_orders::test_refund"]["classification"], "flaky"),
("refund_flips", res["api.test_orders::test_refund"]["flips"], 4),
("refund_flip_rate", res["api.test_orders::test_refund"]["flip_rate"], 1.0),
("refund_fail_rate", res["api.test_orders::test_refund"]["fail_rate"], 0.4),
("refund_duration_spread", res["api.test_orders::test_refund"]["duration_spread"], 10.33),
("error_counts_as_fail", res["api.test_auth::test_expired_token"]["classification"], "always_failing"),
("skip_only_insufficient", res["api.test_search::test_reindex"]["classification"], "insufficient_data"),
("skip_counted", res["api.test_search::test_reindex"]["skips"], 5),
("csv_same_commit_confirmed", res_csv["ui::login"]["classification"], "flaky_confirmed"),
("csv_mixed_commit_listed", res_csv["ui::login"]["mixed_commits"], ["abc"]),
("csv_two_runs_insufficient", res_csv["ui::cart"]["classification"], "insufficient_data"),
("csv_skip_ignored_stable", res_csv["ui::banner"]["classification"], "stable"),
("priority_order", priority(res_csv["ui::login"]) > priority(res["api.test_orders::test_refund"]), True),
("unknown_status_rejected", bad_status, True),
("bad_csv_rejected", bad_csv, True),
("summary_counts", summary(res), {"stable": 1, "flaky": 1, "always_failing": 1, "insufficient_data": 1}),
]
failed = 0
for name, got, expected in cases:
ok = abs(got - expected) < 0.01 if isinstance(expected, float) else got == expected
failed += not ok
print(f"{'PASS' if ok else 'FAIL'} {name}: got={got!r} expected={expected!r}")
print(f"\n{len(cases) - failed}/{len(cases)} passed")
return 1 if failed else 0
def main(argv=None):
p = argparse.ArgumentParser(description="Detect flaky tests from repeated JUnit XML runs or a CSV history.")
p.add_argument("junit", nargs="*", help="JUnit XML files in run order")
p.add_argument("--csv", help="CSV history: run_id,test,status[,commit][,duration_s]")
p.add_argument("--min-runs", type=int, default=3)
p.add_argument("--json", action="store_true", help="print JSON instead of a markdown table")
p.add_argument("--only-flaky", action="store_true")
p.add_argument("--self-test", action="store_true")
a = p.parse_args(argv)
if a.self_test:
return self_test()
rows = []
try:
for i, path in enumerate(a.junit, 1):
with open(path, encoding="utf-8") as fh:
rows += parse_junit_text(fh.read(), f"run{i}:{path}")
if a.csv:
with open(a.csv, encoding="utf-8") as fh:
rows += parse_csv_text(fh.read())
except (OSError, ValueError, ET.ParseError) as e:
print(f"error: {e}", file=sys.stderr)
return 2
if not rows:
p.error("give JUnit XML files, --csv, or --self-test")
res = analyze(rows, a.min_runs)
if a.only_flaky:
res = OrderedDict((k, v) for k, v in res.items() if v["classification"].startswith("flaky"))
if a.json:
print(json.dumps({"summary": summary(res), "tests": res}, indent=2))
else:
print(render_table(res))
print("\nSummary:", ", ".join(f"{k}={v}" for k, v in sorted(summary(res).items())))
return 1 if any(v["classification"].startswith("flaky") for v in res.values()) else 0
if __name__ == "__main__":
sys.exit(main())Scans a codebase for TODO, FIXME, HACK, and XXX comments, extracts owner, ticket, and due date, and flags unowned, untracked, overdue, or vague items with a priority score. Use when cleaning up tech debt, before a release, or asking "which TODOs actually matter?". Includes a stdlib Python scanner with --self-test and a CI-friendly exit code.
---
name: todo-comment-debt-tracker
description: Scans a codebase for TODO, FIXME, HACK, and XXX comments, extracts owner, ticket, and due date, and flags unowned, untracked, overdue, or vague items with a priority score. Use when cleaning up tech debt, before a release, or asking "which TODOs actually matter?". Includes a stdlib Python scanner with --self-test and a CI-friendly exit code.
---
# TODO Comment Debt Tracker
You help teams turn scattered TODO, FIXME, HACK, and XXX comments into a short, owned, prioritized debt list. You scan the code, parse who owns each item, which ticket tracks it, and when it is due, then highlight the ones that are risky or forgotten. You never delete or rewrite comments on your own; you propose changes and let the team decide.
## Files in this skill
- `scripts/todo_debt.py` - scans files or directories, parses owner/ticket/due date, scores each item, and prints a markdown table or JSON (Python 3 standard library only)
- `references/comment-conventions.md` - a recommended TODO format and how to adopt it without churn
- `references/triage-rules.md` - how to decide fix now, ticket, or delete
- `templates/debt-report.md` - report for a sprint review, release checklist, or PR comment
- `examples/sample_service.py` - a small file with typical good and bad TODOs for demos
## Workflow
### 1. Scope the scan
Ask which directories matter and what to skip (generated code, vendored libraries, fixtures). The script already skips `.git`, `node_modules`, `vendor`, `dist`, `build`, virtualenvs, and `target`.
### 2. Run the scanner
```
python3 scripts/todo_debt.py src/ --today 2026-10-18
python3 scripts/todo_debt.py src/ tests/ --ext .py,.ts --json > todo-debt.json
python3 scripts/todo_debt.py examples/sample_service.py --strict
python3 scripts/todo_debt.py --self-test
```
Exit code 1 when an item is overdue or a FIXME/HACK/XXX has no ticket (`--strict`: any problem). Use it as a soft CI gate or a weekly report job.
### 3. Read the problems
- **overdue**: due date passed. Highest priority.
- **no_ticket**: nothing tracks the work; it will be forgotten.
- **stale_hack**: a HACK or XXX with no ticket, often a workaround that outlived its reason.
- **no_owner**: nobody is responsible.
- **vague**: fewer than three words, such as "TODO: fix".
Scores: FIXME/HACK/XXX start at 3, TODO at 1; +2 if overdue, +1 if no ticket.
### 4. Triage with the team
Apply `references/triage-rules.md`: fix now if it is under 30 minutes, create a ticket and add its id if it is real work, delete it if it is obsolete. Use `git blame` on the line to find a likely owner when none is written.
### 5. Report
Fill `templates/debt-report.md` with the top items, totals by marker and owner, and agreed actions. Suggest the convention in `references/comment-conventions.md` for new comments.
## Rules
- Do not count markers inside string literals or plain prose; only comments.
- Never invent ticket numbers or owners; mark them unknown and suggest how to find them.
- Do not remove comments in bulk without review; propose a patch instead.
- Respect generated and vendored code: report it separately or skip it.
FILE:references/comment-conventions.md
# Recommended TODO convention
```
# TODO(owner, TICKET-123, 2026-12-01): what needs to happen and why
```
- **owner**: a username or team handle (`alice`, `@payments-team`). One name, the person who will make the call.
- **ticket**: `#123`, `PROJ-123`, or a full issue URL. Required for FIXME, HACK, and XXX.
- **due date**: ISO `YYYY-MM-DD`, optional for TODO, recommended for HACK workarounds.
- **text**: an action plus the reason, at least a short sentence. "TODO: fix" is not useful.
## Which marker to use
| Marker | Meaning | Ticket required? |
|---|---|---|
| TODO | Planned improvement, not wrong today | Recommended |
| FIXME | Known bug or incorrect behavior | Yes |
| HACK | Deliberate workaround that should be removed | Yes, plus a due date |
| XXX | Danger, surprising code, review needed | Yes |
## Adopting without churn
1. Run the scanner and publish the report; do not mass-edit.
2. Apply the convention to new comments via code review or a lint rule.
3. Fix old comments when you touch that code anyway.
4. Revisit the totals monthly; aim for "no overdue, no untracked FIXME".
FILE:references/triage-rules.md
# Triage rules
For each item, ask in order:
1. **Is it still true?** Read the surrounding code and `git log -L` for the lines. If the issue is gone, delete the comment.
2. **Is it under 30 minutes?** Fix it now in a small PR that removes the comment.
3. **Is it a bug or a risk?** (FIXME/XXX, security, data loss, money) Create a ticket with priority, add the ticket id to the comment, and assign an owner.
4. **Is it a workaround?** (HACK) Write down the condition that allows removal ("remove after vendor SDK 5.x") and a due date.
5. **Is it a wish?** If nobody would schedule it in the next two quarters, delete it or move it to a backlog ticket.
## Priority hints
- Overdue items first, then FIXME/HACK/XXX without tickets, then unowned items in hot files (high churn).
- Items in auth, payments, data migrations, or concurrency code deserve a closer look regardless of score.
## Finding an owner
- `git blame -L <line>,<line> <file>` shows who added it; that person or their team is the default owner.
- CODEOWNERS files map paths to teams.
- If the author has left, assign to the current maintainer of the module.
FILE:templates/debt-report.md
# TODO Debt Report
**Repository:** <name> **Scanned:** <paths> **Date:** <YYYY-MM-DD>
## Totals
| Marker | Count | Overdue | No ticket | No owner |
|---|---|---|---|---|
| TODO | | | | |
| FIXME | | | | |
| HACK | | | | |
| XXX | | | | |
## Top 10 by score
| Score | Location | Marker | Owner | Ticket | Due | Problems | Proposed action |
|---|---|---|---|---|---|---|---|
| | | | | | | | |
## By owner
| Owner | Items | Overdue |
|---|---|---|
## Agreed actions
- [ ] Fix now:
- [ ] Ticket created:
- [ ] Delete (obsolete):
## Convention reminder
`# TODO(owner, TICKET-123, YYYY-MM-DD): action and reason`
FILE:examples/sample_service.py
"""Sample order service with typical TODO comments for demos of todo_debt.py."""
import json
TAX_RATE = 0.2 # XXX(PAY-311): confirm tax rate per region with finance
def load_orders(path):
# TODO(maria, 2026-03-01): switch to streaming parser once files exceed 50 MB (#88)
with open(path, encoding="utf-8") as fh:
return json.load(fh)
def total(order):
# FIXME: returns wrong total when a discount exceeds the subtotal
subtotal = sum(item["price"] * item["qty"] for item in order["items"])
discount = order.get("discount", 0)
return round((subtotal - discount) * (1 + TAX_RATE), 2)
def notify(order):
# HACK [2026-11-30] skip emails for test accounts until the staging mail stub is fixed
if order.get("email", "").endswith("@test.local"):
return False
# TODO: retry
return True
def label(order):
note = "TODO: this string is not a comment and should be ignored"
# TODO(@ken, OPS-19): move label printing to the warehouse service
return f"{order['id']} {note[:0]}"
FILE:scripts/todo_debt.py
#!/usr/bin/env python3
"""Scan source files for TODO/FIXME/HACK/XXX comments and grade them (Python 3 standard library only).
Recognized forms (marker must follow a comment token: # // /* * -- ; <!--):
# TODO: plain note
// TODO(alice): owner
# FIXME(#123): issue number # TODO(PROJ-42) ticket key # TODO(https://...) link
# TODO(bob, 2026-12-01): owner + due date # HACK [2026-11-15] due date in brackets
Problems flagged per item:
no_owner, no_ticket, overdue (due date before --today), stale_hack (HACK/XXX without ticket),
vague (fewer than 3 words of text)
Exit code: 1 if any item is overdue or a FIXME/HACK/XXX has no ticket (with --strict: any problem).
"""
import argparse
import datetime as dt
import json
import os
import re
import sys
MARKERS = ("TODO", "FIXME", "HACK", "XXX")
SEVERITY = {"FIXME": 3, "HACK": 3, "XXX": 3, "TODO": 1}
COMMENT_RE = re.compile(
r"(?:#|//|/\*|(?:^|\s)\*|--|;|<!--)\s*(?P<marker>TODO|FIXME|HACK|XXX)\b"
r"(?:\s*\((?P<paren>[^)]*)\))?(?:\s*\[(?P<bracket>[^\]]*)\])?\s*:?\s*(?P<text>.*)$")
DATE_RE = re.compile(r"\b(\d{4}-\d{2}-\d{2})\b")
TICKET_RE = re.compile(r"(#\d+\b|\b[A-Z][A-Z0-9]+-\d+\b|https?://\S+)")
OWNER_RE = re.compile(r"^@?([A-Za-z][\w.-]*)$")
DEFAULT_EXT = {".py", ".js", ".ts", ".tsx", ".jsx", ".go", ".rs", ".java", ".kt", ".c", ".h", ".cpp",
".cs", ".rb", ".php", ".swift", ".sh", ".sql", ".yaml", ".yml", ".toml", ".css", ".scss",
".html", ".vue", ".lua"}
SKIP_DIRS = {".git", "node_modules", "vendor", "dist", "build", ".venv", "venv", "__pycache__", "target"}
def parse_line(line, path="<text>", lineno=0, today=None):
m = COMMENT_RE.search(line)
if not m:
return None
marker = m.group("marker")
meta = ",".join(x for x in (m.group("paren"), m.group("bracket")) if x)
text = (m.group("text") or "").strip().rstrip("*/").rstrip("-->").strip()
owner = None
ticket = None
due = None
for part in [p.strip() for p in meta.split(",") if p.strip()]:
if DATE_RE.fullmatch(part.replace("by ", "").replace("until ", "").strip()):
due = DATE_RE.search(part).group(1)
elif TICKET_RE.fullmatch(part):
ticket = part
elif OWNER_RE.fullmatch(part) and owner is None:
owner = OWNER_RE.fullmatch(part).group(1)
if ticket is None:
t = TICKET_RE.search(text)
if t:
ticket = t.group(1)
if due is None:
d = re.search(r"\b(?:by|until|before|due)\s+(\d{4}-\d{2}-\d{2})\b", text)
if d:
due = d.group(1)
problems = []
if not owner:
problems.append("no_owner")
if not ticket:
problems.append("no_ticket")
if marker in ("HACK", "XXX"):
problems.append("stale_hack")
if due and today:
try:
if dt.date.fromisoformat(due) < today:
problems.append("overdue")
except ValueError:
problems.append("bad_date")
if len(re.findall(r"[A-Za-z]{2,}", text)) < 3:
problems.append("vague")
return {"file": path, "line": lineno, "marker": marker, "owner": owner, "ticket": ticket,
"due": due, "text": text, "problems": problems,
"score": SEVERITY[marker] + 2 * ("overdue" in problems) + ("no_ticket" in problems)}
def scan_text(text, path="<text>", today=None):
items = []
for i, line in enumerate(text.splitlines(), 1):
it = parse_line(line, path, i, today)
if it:
items.append(it)
return items
def iter_files(root, exts):
if os.path.isfile(root):
yield root
return
for dirpath, dirnames, filenames in os.walk(root):
dirnames[:] = sorted(d for d in dirnames if d not in SKIP_DIRS)
for fn in sorted(filenames):
if os.path.splitext(fn)[1].lower() in exts:
yield os.path.join(dirpath, fn)
def is_blocking(item, strict=False):
if strict:
return bool(item["problems"])
return "overdue" in item["problems"] or (item["marker"] != "TODO" and "no_ticket" in item["problems"])
def summarize(items):
s = {"total": len(items), "by_marker": {}, "by_owner": {}, "problems": {}}
for it in items:
s["by_marker"][it["marker"]] = s["by_marker"].get(it["marker"], 0) + 1
o = it["owner"] or "(unowned)"
s["by_owner"][o] = s["by_owner"].get(o, 0) + 1
for p in it["problems"]:
s["problems"][p] = s["problems"].get(p, 0) + 1
return s
def render(items):
out = ["| Score | Marker | Location | Owner | Ticket | Due | Problems | Text |", "|---|---|---|---|---|---|---|---|"]
for it in sorted(items, key=lambda x: (-x["score"], x["file"], x["line"])):
out.append(f"| {it['score']} | {it['marker']} | {it['file']}:{it['line']} | {it['owner'] or '-'} | "
f"{it['ticket'] or '-'} | {it['due'] or '-'} | {', '.join(it['problems']) or 'ok'} | {it['text'][:60]} |")
return "\n".join(out)
SAMPLE = '''import os
# TODO(alice, 2026-01-15): remove legacy CSV export after #412 ships
def f():
pass # FIXME: crashes when list is empty
// HACK [2026-12-31] skip auth in local dev
/* XXX(PROJ-77): double-check rounding with finance */
-- TODO(@bob): add index
x = "TODO: this is a string, not a comment"
<!-- TODO(carol): https://example.com/issues/9 replace hero image -->
* TODO update docs by 2025-05-01
# todo lowercase is ignored
'''
def self_test():
today = dt.date(2026, 10, 18)
items = scan_text(SAMPLE, "sample", today)
by_line = {it["line"]: it for it in items}
cases = [
("count", len(items), 7),
("owner_parsed", by_line[2]["owner"], "alice"),
("due_parsed", by_line[2]["due"], "2026-01-15"),
("overdue_flag", "overdue" in by_line[2]["problems"], True),
("ticket_from_text", by_line[2]["ticket"], "#412"),
("fixme_no_ticket", by_line[4]["problems"][:2], ["no_owner", "no_ticket"]),
("fixme_blocking", is_blocking(by_line[4]), True),
("hack_bracket_due_future", "overdue" in by_line[5]["problems"], False),
("hack_stale", "stale_hack" in by_line[5]["problems"], True),
("xxx_ticket_key", by_line[6]["ticket"], "PROJ-77"),
("at_owner", by_line[7]["owner"], "bob"),
("vague_short", "vague" in by_line[7]["problems"], True),
("string_not_comment", 8 in by_line, False),
("html_comment_link", by_line[9]["ticket"], "https://example.com/issues/9"),
("star_comment_due_text", by_line[10]["due"], "2025-05-01"),
("lowercase_ignored", 11 in by_line, False),
("todo_not_blocking_default", is_blocking(by_line[7]), False),
("todo_blocking_strict", is_blocking(by_line[7], strict=True), True),
("summary_markers", summarize(items)["by_marker"], {"TODO": 4, "FIXME": 1, "HACK": 1, "XXX": 1}),
("score_overdue_todo", by_line[2]["score"], 3),
]
failed = 0
for name, got, expected in cases:
ok = got == expected
failed += not ok
print(f"{'PASS' if ok else 'FAIL'} {name}: got={got!r} expected={expected!r}")
print(f"\n{len(cases) - failed}/{len(cases)} passed")
return 1 if failed else 0
def main(argv=None):
p = argparse.ArgumentParser(description="Find and grade TODO/FIXME/HACK/XXX comments.")
p.add_argument("paths", nargs="*", default=[], help="files or directories to scan")
p.add_argument("--today", help="YYYY-MM-DD for overdue checks (default: today)")
p.add_argument("--ext", help="comma-separated extensions to include, e.g. .py,.ts")
p.add_argument("--json", action="store_true")
p.add_argument("--strict", action="store_true", help="exit 1 on any problem")
p.add_argument("--self-test", action="store_true")
a = p.parse_args(argv)
if a.self_test:
return self_test()
if not a.paths:
p.error("give files/directories to scan, or --self-test")
today = dt.date.fromisoformat(a.today) if a.today else dt.date.today()
exts = {e if e.startswith(".") else "." + e for e in a.ext.split(",")} if a.ext else DEFAULT_EXT
items = []
for root in a.paths:
for path in iter_files(root, exts):
try:
with open(path, encoding="utf-8", errors="replace") as fh:
items += scan_text(fh.read(), path, today)
except OSError as e:
print(f"warn: {path}: {e}", file=sys.stderr)
if a.json:
print(json.dumps({"summary": summarize(items), "items": items}, indent=2))
else:
print(render(items))
s = summarize(items)
print(f"\nTotal {s['total']}; markers {s['by_marker']}; problems {s['problems']}")
return 1 if any(is_blocking(it, a.strict) for it in items) else 0
if __name__ == "__main__":
sys.exit(main())Plans a fair, rotating weekly chore schedule for a household: weighs each chore by minutes and frequency, respects capacity shares and "cannot do" limits, rotates jobs week to week, and reports each person's load against a fair target. Use when roommates or families argue about chores or ask "who does what this month?". Includes a stdlib Python planner with --self-test.
---
name: household-chore-rotation-planner
description: Plans a fair, rotating weekly chore schedule for a household: weighs each chore by minutes and frequency, respects capacity shares and "cannot do" limits, rotates jobs week to week, and reports each person's load against a fair target. Use when roommates or families argue about chores or ask "who does what this month?". Includes a stdlib Python planner with --self-test.
---
# Household Chore Rotation Planner
You help households (roommates, couples, families with kids) split chores fairly and rotate them so nobody is stuck with the worst jobs forever. You measure fairness in minutes, not in number of tasks, adjust for each person's capacity, respect real limits (age, allergies, health, skills), and produce a schedule people can stick on the fridge. You stay neutral and kind; you never take sides in a household argument.
## Files in this skill
- `scripts/chore_rotation.py` - builds a multi-week rotation from a JSON list of people and chores and reports minutes vs fair target per person (Python 3 standard library only)
- `references/chore-time-estimates.md` - typical minutes for common chores to use as defaults
- `references/fairness-and-agreements.md` - how to set shares, handle swaps, and run a short house meeting
- `templates/rotation-chart.md` - printable weekly chart plus a house agreement section
- `examples/household.json` - four-person sample household with constraints
## Workflow
### 1. Gather the household
Ask for:
- People and their capacity: full share (1.0), reduced share (for example 0.5 for a young child or someone working night shifts), or larger share by agreement
- Hard limits: chores someone cannot do (`cannot`), or chores only certain people can do (`only`, for example cooking or driving)
- Chores with realistic minutes per session and times per week; use `references/chore-time-estimates.md` when unsure
- How many weeks to plan (4 is a good default) and any preferences ("I would rather cook than clean the bathroom")
### 2. Build the input and run the planner
```
python3 scripts/chore_rotation.py examples/household.json
python3 scripts/chore_rotation.py my-house.json --weeks 6 --json
python3 scripts/chore_rotation.py --self-test
```
Exit code 1 means a chore has no eligible person (fix the `cannot`/`only` lists); 2 means bad input.
### 3. Check fairness
Read the "Overall fairness" table. Within about 10% of target is good. If someone is far off, adjust minutes, split a big chore (for example "Laundry - wash" and "Laundry - fold"), or change shares by agreement, then rerun.
### 4. Apply preferences
The planner balances minutes; it does not know who hates what. Offer swaps of similar minutes between people, and keep the totals close. If a heavy chore lands on the same person every week because of constraints, say so and suggest splitting it.
### 5. Deliver
Fill `templates/rotation-chart.md` with the weekly tables, a short house agreement (swap rules, what "done" means, a monthly check-in), and a friendly group message under 80 words announcing the plan.
## Rules
- Fairness is measured in minutes weighted by agreed shares, not task counts.
- Never assign unsafe chores to children (sharp tools, strong chemicals, ladders, hot ovens without supervision).
- Keep the tone neutral; describe loads with numbers, not blame.
- Invisible work counts: planning meals, booking repairs, and buying supplies can be added as chores with minutes.
FILE:references/chore-time-estimates.md
# Typical chore times (defaults, adjust to your home)
| Chore | Minutes per session | Common frequency |
|---|---|---|
| Dishes / load and unload dishwasher | 15-25 | daily |
| Cook dinner | 30-60 | 3-7 per week |
| Meal planning and grocery list | 20-30 | weekly |
| Grocery run | 45-75 | weekly |
| Laundry (wash, dry, fold, put away) | 40-60 per load | 2-4 loads per week |
| Bathroom clean (sink, toilet, shower, floor) | 30-45 | weekly |
| Kitchen deep clean (counters, stovetop, fridge wipe) | 30-40 | weekly |
| Vacuum and mop | 30-45 | 1-2 per week |
| Dusting | 15-20 | weekly |
| Trash and recycling out | 5-10 | 2-3 per week |
| Change bed linens | 15 per bed | every 1-2 weeks |
| Water plants | 5-15 | 2 per week |
| Pet feeding | 5 | daily, often twice |
| Litter box / pet cleanup | 5 | daily |
| Dog walk | 20-30 | 1-3 per day |
| Mow lawn | 40-60 | weekly in season |
| Clean oven | 30-45 | monthly |
| Pay bills and admin | 20-30 | weekly |
Tips
- Time one session with a phone timer if people disagree about how long a chore takes.
- Split chores that are both long and unpopular so they can rotate more easily.
- Seasonal chores (snow shoveling, gutters) can be added only for the weeks they apply.
FILE:references/fairness-and-agreements.md
# Fairness and house agreements
## Setting shares
- Start everyone at 1.0. Lower shares only by explicit agreement (age, health, very long work hours, exams).
- Children: roughly 0.25-0.5 with age-appropriate chores only.
- Someone who pays more rent or does more invisible work may get a lower chore share; write it down.
## What "done" means
Agree on a one-line definition per chore, for example "Dishes: sink empty, counters wiped, dishwasher running".
## Swaps
- Swaps are fine if both people agree and minutes are similar.
- Post swaps in the group chat so nobody is surprised.
- A missed chore is made up within 48 hours or swapped, not silently skipped.
## 15-minute monthly check-in
1. Each person: one thing working, one thing not (2 minutes each).
2. Look at the fairness table; adjust minutes or shares if reality changed.
3. Decide on one improvement for next month.
4. Regenerate the rotation.
## Handling conflict
- Talk about the system, not the person ("Bathroom keeps slipping on Thursdays" instead of "You never clean").
- If one person carries far more invisible work, add it to the chore list with minutes.
FILE:templates/rotation-chart.md
# Chore Rotation - <household name>
**Period:** <start date> to <end date> **People:** <names and shares>
## Week 1 (<dates>)
| Chore | Who | Done? |
|---|---|---|
| | | [ ] |
## Week 2 (<dates>)
| Chore | Who | Done? |
|---|---|---|
| | | [ ] |
## Fairness (minutes)
| Person | Share | Minutes | Target | Deviation |
|---|---|---|---|---|
## House agreement
- "Done" means:
- Swap rule:
- Missed chore rule:
- Monthly check-in date:
## Group message
> Hi all! New chore rotation for <period> is on the fridge and pinned here. Loads are balanced by time. Swaps are fine, just post them. Check-in on <date>.
FILE:examples/household.json
{
"weeks": 4,
"people": [
{"name": "Leyla", "share": 1.0},
{"name": "Tom", "share": 1.0, "cannot": ["Litter box"]},
{"name": "Deniz", "share": 0.5, "cannot": ["Clean oven", "Mow lawn"]},
{"name": "Sam", "share": 0.75}
],
"chores": [
{"name": "Dishes", "minutes": 20, "times_per_week": 7},
{"name": "Cook dinner", "minutes": 45, "times_per_week": 3, "only": ["Leyla", "Tom", "Sam"]},
{"name": "Laundry", "minutes": 40, "times_per_week": 2},
{"name": "Bathroom clean", "minutes": 40, "times_per_week": 1},
{"name": "Vacuum and mop", "minutes": 35, "times_per_week": 2},
{"name": "Trash and recycling", "minutes": 10, "times_per_week": 3},
{"name": "Grocery run", "minutes": 60, "times_per_week": 1},
{"name": "Litter box", "minutes": 5, "times_per_week": 7},
{"name": "Water plants", "minutes": 10, "times_per_week": 2},
{"name": "Mow lawn", "minutes": 50, "times_per_week": 1},
{"name": "Clean oven", "minutes": 30, "times_per_week": 1}
]
}
FILE:scripts/chore_rotation.py
#!/usr/bin/env python3
"""Fair household chore rotation planner (Python 3 standard library only).
Input JSON:
{
"weeks": 4,
"people": [{"name": "Ana", "share": 1.0, "cannot": ["Mow lawn"]}, ...],
"chores": [{"name": "Dishes", "minutes": 20, "times_per_week": 7, "only": ["Ana", "Ben"]}, ...]
}
share: relative capacity (1.0 = normal; 0.5 = half load, e.g. a child or someone with long shifts)
cannot / only: hard constraints (allergies, age, physical limits, skills)
Algorithm per week: chores sorted by weekly minutes (largest first) go to the eligible person with the
lowest (load / share); ties are broken in favor of someone who did NOT have that chore last week,
so chores rotate. Reports per-person minutes, target minutes, and deviation.
Exit code: 0 OK, 1 if a chore has no eligible person, 2 on bad input.
"""
import argparse
import json
import sys
class PlanError(ValueError):
pass
def validate(data):
people = data.get("people") or []
chores = data.get("chores") or []
if not people:
raise PlanError("need at least one person")
if not chores:
raise PlanError("need at least one chore")
names = [p["name"] for p in people]
if len(set(names)) != len(names):
raise PlanError("duplicate person names")
for p in people:
if float(p.get("share", 1.0)) <= 0:
raise PlanError(f"share must be > 0 for {p['name']}")
for c in chores:
if float(c.get("minutes", 0)) <= 0 or float(c.get("times_per_week", 1)) <= 0:
raise PlanError(f"minutes and times_per_week must be > 0 for {c.get('name')}")
for n in c.get("only", []):
if n not in names:
raise PlanError(f"chore {c['name']} lists unknown person {n}")
return people, chores
def weekly_minutes(chore):
return float(chore["minutes"]) * float(chore.get("times_per_week", 1))
def eligible(person, chore):
if chore.get("only") and person["name"] not in chore["only"]:
return False
return chore["name"] not in person.get("cannot", [])
def plan_week(people, chores, last_week=None):
"""Return (assignment {chore: person}, loads {person: minutes}, unassigned [chore])."""
last_week = last_week or {}
loads = {p["name"]: 0.0 for p in people}
assignment = {}
unassigned = []
order = sorted(chores, key=lambda c: (-weekly_minutes(c), c["name"]))
for c in order:
cands = [p for p in people if eligible(p, c)]
if not cands:
unassigned.append(c["name"])
continue
w = weekly_minutes(c)
def key(p):
share = float(p.get("share", 1.0))
after = (loads[p["name"]] + w) / share
repeat = 1 if last_week.get(c["name"]) == p["name"] else 0
return (round(after, 6), repeat, p["name"])
# prefer rotation when loads are close: compare within a 10% band of the best option
ranked = sorted(cands, key=key)
best = key(ranked[0])[0]
band = [p for p in ranked if key(p)[0] <= best * 1.10 + 1e-9]
band.sort(key=lambda p: (key(p)[1], key(p)[0], p["name"]))
chosen = band[0]
assignment[c["name"]] = chosen["name"]
loads[chosen["name"]] += w
return assignment, loads, unassigned
def fairness(people, loads):
total = sum(loads.values())
total_share = sum(float(p.get("share", 1.0)) for p in people)
rows = []
for p in people:
target = total * float(p.get("share", 1.0)) / total_share
got = loads[p["name"]]
dev = (got - target) / target * 100 if target else 0.0
rows.append({"name": p["name"], "minutes": round(got, 1), "target": round(target, 1),
"deviation_pct": round(dev, 1)})
return rows
def plan(data):
people, chores = validate(data)
weeks = int(data.get("weeks", 4))
result = []
cumulative = {p["name"]: 0.0 for p in people}
last = None
for wk in range(1, weeks + 1):
assignment, loads, unassigned = plan_week(people, chores, last)
for k, v in loads.items():
cumulative[k] += v
result.append({"week": wk, "assignment": assignment, "fairness": fairness(people, loads),
"unassigned": unassigned})
last = assignment
return {"weeks": result, "overall": fairness(people, cumulative)}
def render(res):
out = []
for w in res["weeks"]:
out.append(f"## Week {w['week']}")
out.append("| Chore | Who |")
out.append("|---|---|")
for chore, who in sorted(w["assignment"].items()):
out.append(f"| {chore} | {who} |")
for u in w["unassigned"]:
out.append(f"| {u} | UNASSIGNED (no eligible person) |")
out.append("")
out.append("## Overall fairness (minutes across all weeks)")
out.append("| Person | Minutes | Target | Deviation |")
out.append("|---|---|---|---|")
for r in res["overall"]:
out.append(f"| {r['name']} | {r['minutes']} | {r['target']} | {r['deviation_pct']:+.1f}% |")
return "\n".join(out)
def self_test():
base = {
"weeks": 4,
"people": [{"name": "Ana"}, {"name": "Ben"}, {"name": "Cem", "share": 0.5, "cannot": ["Mow lawn"]}],
"chores": [
{"name": "Dishes", "minutes": 20, "times_per_week": 7},
{"name": "Laundry", "minutes": 60, "times_per_week": 2},
{"name": "Bathroom", "minutes": 45, "times_per_week": 1},
{"name": "Vacuum", "minutes": 30, "times_per_week": 2},
{"name": "Trash", "minutes": 5, "times_per_week": 3},
{"name": "Mow lawn", "minutes": 50, "times_per_week": 1},
{"name": "Feed cat", "minutes": 5, "times_per_week": 14, "only": ["Cem", "Ana"]},
],
}
res = plan(base)
w1 = res["weeks"][0]["assignment"]
lawn_never_cem = all(w["assignment"]["Mow lawn"] != "Cem" for w in res["weeks"])
cat_ok = all(w["assignment"]["Feed cat"] in ("Cem", "Ana") for w in res["weeks"])
dishes_people = {w["assignment"]["Dishes"] for w in res["weeks"]}
overall = {r["name"]: r for r in res["overall"]}
max_dev = max(abs(r["deviation_pct"]) for r in res["overall"])
two = plan({"weeks": 2, "people": [{"name": "A"}, {"name": "B"}],
"chores": [{"name": "X", "minutes": 10}, {"name": "Y", "minutes": 10}]})
rotated = two["weeks"][0]["assignment"]["X"] != two["weeks"][1]["assignment"]["X"]
orphan = plan({"weeks": 1, "people": [{"name": "A", "cannot": ["Z"]}], "chores": [{"name": "Z", "minutes": 5}]})
errors = []
for bad in ({"people": [], "chores": [{"name": "x", "minutes": 1}]},
{"people": [{"name": "A"}, {"name": "A"}], "chores": [{"name": "x", "minutes": 1}]},
{"people": [{"name": "A", "share": 0}], "chores": [{"name": "x", "minutes": 1}]},
{"people": [{"name": "A"}], "chores": [{"name": "x", "minutes": 0}]},
{"people": [{"name": "A"}], "chores": [{"name": "x", "minutes": 1, "only": ["Q"]}]}):
try:
plan(bad)
errors.append(False)
except PlanError:
errors.append(True)
cases = [
("all_chores_assigned_w1", len(w1), 7),
("weekly_minutes", weekly_minutes(base["chores"][0]), 140.0),
("cannot_respected", lawn_never_cem, True),
("only_respected", cat_ok, True),
("dishes_rotate", len(dishes_people) >= 2, True),
("half_share_lower_load", overall["Cem"]["minutes"] < overall["Ana"]["minutes"], True),
("overall_fair_within_25pct", max_dev <= 25.0, True),
("targets_sum_to_total", round(sum(r["target"] for r in res["overall"]), 0) == round(sum(r["minutes"] for r in res["overall"]), 0), True),
("equal_tie_rotates", rotated, True),
("orphan_reported", orphan["weeks"][0]["unassigned"], ["Z"]),
("bad_inputs_rejected", errors, [True] * 5),
("render_has_header", render(res).startswith("## Week 1"), True),
]
failed = 0
for name, got, expected in cases:
ok = got == expected
failed += not ok
print(f"{'PASS' if ok else 'FAIL'} {name}: got={got!r} expected={expected!r}")
print(f"\n{len(cases) - failed}/{len(cases)} passed")
return 1 if failed else 0
def main(argv=None):
p = argparse.ArgumentParser(description="Plan a fair, rotating household chore schedule.")
p.add_argument("input", nargs="?", help="JSON file with weeks, people, chores")
p.add_argument("--weeks", type=int, help="override number of weeks")
p.add_argument("--json", action="store_true")
p.add_argument("--self-test", action="store_true")
a = p.parse_args(argv)
if a.self_test:
return self_test()
if not a.input:
p.error("give an input JSON file, or --self-test")
try:
with open(a.input, encoding="utf-8") as fh:
data = json.load(fh)
if a.weeks:
data["weeks"] = a.weeks
res = plan(data)
except (OSError, json.JSONDecodeError, PlanError, KeyError) as e:
print(f"error: {e}", file=sys.stderr)
return 2
print(json.dumps(res, indent=2) if a.json else render(res))
return 1 if any(w["unassigned"] for w in res["weeks"]) else 0
if __name__ == "__main__":
sys.exit(main())Decodes a JWT without trusting it and lints the header and claims: alg none, remote jku/x5u keys, missing or expired exp, nbf/iat clock skew, millisecond timestamps, long lifetimes, iss/aud mismatches, sensitive data in the payload, and HMAC signature checks with a secret from an env var. Use when a login or API call fails with 401, or when reviewing token settings. Includes a stdlib Python inspector with --self-test.
---
name: jwt-token-inspector
description: Decodes a JWT without trusting it and lints the header and claims: alg none, remote jku/x5u keys, missing or expired exp, nbf/iat clock skew, millisecond timestamps, long lifetimes, iss/aud mismatches, sensitive data in the payload, and HMAC signature checks with a secret from an env var. Use when a login or API call fails with 401, or when reviewing token settings. Includes a stdlib Python inspector with --self-test.
---
# JWT Token Inspector
You help developers understand and debug JSON Web Tokens safely. You decode the header and payload, translate timestamps into readable UTC times, explain what each claim means, and flag settings that make the token unsafe, unusable, or likely to break in production. You treat every token as untrusted: decoding is not verification, and you say so.
## Files in this skill
- `scripts/jwt_inspect.py` - decodes a JWS compact token, lints header and claims, optionally verifies HS256/HS384/HS512 with a secret, prints markdown or JSON (Python 3 standard library only)
- `references/claims-and-validation.md` - registered claims, what a verifier must check, and in which order
- `references/common-jwt-pitfalls.md` - the mistakes behind most JWT bugs and incidents, with fixes
- `templates/jwt-review.md` - a short report for a bug ticket, PR review, or security review
- `examples/sample-tokens.md` - three demo tokens (long-lived, alg none, remote jku with millisecond exp) and expected findings
## Workflow
### 1. Handle the token safely
Ask for a token from a test or staging environment. If the user pastes a live production token, warn them that it grants access until it expires and suggest rotating it after the session. Never repeat the full token or signature back in your answer.
### 2. Run the inspector
```
python3 scripts/jwt_inspect.py "<token>"
python3 scripts/jwt_inspect.py "<token>" --now 2026-10-19T09:00:00Z --leeway 60
python3 scripts/jwt_inspect.py "<token>" --expect-iss https://auth.example.com --expect-aud api://orders
JWT_SECRET=... python3 scripts/jwt_inspect.py "<token>" --secret-env JWT_SECRET
python3 scripts/jwt_inspect.py --self-test
```
A `Bearer ` prefix is stripped. Use `-` to read the token from stdin. Exit code 1 means at least one high finding (`--strict`: any finding); 2 means the token could not be decoded.
### 3. Explain the result
- Show the decoded header and claims and the exp/nbf/iat times in UTC and in the user's time zone.
- Group findings by severity: **high** (unsafe or rejected: alg none, empty signature, expired, not yet valid, iss/aud mismatch, remote or embedded keys, bad signature, millisecond timestamps), **medium** (no exp, lifetime over policy, sensitive claims, weak HMAC secret, iat in the future), **low** (missing iss, aud, sub, or kid).
- For a 401, match the finding to the symptom using `references/claims-and-validation.md`: clock skew, wrong audience, wrong issuer, or wrong key are the usual causes.
### 4. Recommend fixes
Use `references/common-jwt-pitfalls.md`: pin the algorithm list on the verifier, keep access tokens short (5-15 minutes) with refresh tokens, move personal data out of the payload, allow 30-60 seconds of leeway, and fetch keys only from a configured JWKS URL.
### 5. Report
Fill `templates/jwt-review.md` with the token fingerprint (never the token), findings, root cause, and the agreed fix.
## Rules
- Decoding is not verification. Never say a token is valid unless its signature was checked with the right key.
- Never print or store full tokens, signatures, or secrets. Pass secrets through an environment variable, not the command line.
- Do not help forge, replay, or tamper with tokens for systems the user does not own.
- RS/ES/PS/EdDSA signatures need the public key and a proper JOSE library; the script only verifies HMAC algorithms.
FILE:references/claims-and-validation.md
# Claims and Validation Order
## Registered claims (RFC 7519)
| Claim | Meaning | Type | Notes |
|---|---|---|---|
| iss | Issuer | string/URL | Pin it on the verifier; one issuer per trust domain |
| sub | Subject (user or service id) | string | Stable id, not an email that can change |
| aud | Audience | string or array | The API must find its own id in it |
| exp | Expiration time | NumericDate (seconds) | Required for access tokens |
| nbf | Not before | NumericDate | Token is rejected until then |
| iat | Issued at | NumericDate | Used for lifetime and freshness checks |
| jti | Token id | string | Enables revocation lists and replay detection |
NumericDate is seconds since 1970-01-01T00:00:00Z. A 13-digit value is milliseconds and will be read as a date tens of thousands of years away.
## Header parameters
- `alg` - signing algorithm. The verifier decides which values are allowed; the token never does.
- `kid` - key id, used to pick the right key from the issuer's JWKS during rotation.
- `typ` - usually `JWT` or `at+jwt` (access token profile, RFC 9068).
- `jku`, `x5u`, `jwk` - point to or embed keys. Ignore them unless they match a pinned, trusted source.
## What a verifier must check, in order
1. Token has three segments and decodes to JSON objects.
2. `alg` is in the verifier's allow-list (for example only RS256). Reject `none`.
3. Signature verifies with a key chosen by `kid` from a trusted JWKS, never from the token itself.
4. `exp` is in the future and `nbf` is in the past, both with small leeway (30-60 s).
5. `iss` equals the expected issuer.
6. `aud` contains this service's id.
7. Application rules: scopes, roles, tenant, `jti` not revoked.
## Matching 401 symptoms to causes
| Symptom | Likely cause | Check |
|---|---|---|
| Works locally, fails in prod | Clock skew or different issuer URL | iat/nbf vs server time, iss |
| Fails right after login | nbf or iat in the future | server clocks (NTP) |
| Fails for one API only | Wrong audience | aud list |
| Fails after key rotation | Unknown kid or cached JWKS | kid vs JWKS, cache TTL |
| Never expires | exp missing or in milliseconds | exp value |
FILE:references/common-jwt-pitfalls.md
# Common JWT Pitfalls
1. **Accepting `alg: none`.** Unsigned tokens are forgeable. Fix: pass an explicit algorithm allow-list to the verify call.
2. **Algorithm confusion (RS256 vs HS256).** If a verifier accepts both, an attacker can sign with HMAC using the public key as the secret. Fix: one algorithm family per key.
3. **Trusting keys from the token.** `jku`, `x5u`, or an embedded `jwk` let the sender choose the key. Fix: load keys only from a configured JWKS URL.
4. **Weak HMAC secrets.** Short or dictionary secrets can be brute-forced offline from one token. Fix: at least 32 random bytes for HS256 (48 for HS384, 64 for HS512), stored in a secret manager.
5. **Long-lived access tokens.** A leaked token works until exp. Fix: 5-15 minute access tokens plus rotating refresh tokens; revoke by `jti` if needed.
6. **Personal data in the payload.** Base64url is encoding, not encryption; anyone holding the token can read it. Fix: keep ids and scopes only, or use JWE.
7. **Milliseconds instead of seconds.** Common with `Date.now()` in JavaScript. Fix: `Math.floor(Date.now() / 1000)`.
8. **No leeway.** Small clock differences cause random 401s. Fix: 30-60 seconds of leeway and NTP on all hosts.
9. **Skipping aud/iss checks.** A token for one service gets replayed against another. Fix: always validate both.
10. **Tokens in URLs or logs.** They leak through referrers, proxies, and log search. Fix: send in the Authorization header and redact in logs.
FILE:templates/jwt-review.md
# JWT Review - <service or ticket>
**Token fingerprint:** sha256 <first 12 hex chars> **Environment:** <dev / staging / prod> **Checked at:** <UTC time>
**Issuer:** <iss> **Audience:** <aud> **Algorithm:** <alg> **Lifetime:** <exp - iat>
## Findings
| Severity | Code | Detail | Fix |
|---|---|---|---|
| | | | |
## Root cause
<one or two sentences tying the finding to the reported symptom>
## Agreed changes
- [ ] Verifier: <allow-list, leeway, iss/aud checks>
- [ ] Issuer: <lifetime, claims, key size, kid>
- [ ] Rotate exposed token or secret: <yes/no, owner>
## Notes
Decoded with scripts/jwt_inspect.py. The token itself is not stored in this report.
FILE:examples/sample-tokens.md
# Sample Tokens (demo only, no real systems)
Run each with:
```
python3 scripts/jwt_inspect.py "<token>" --now 2026-10-19T09:00:00Z --expect-aud api://orders
```
## 1. Long-lived customer token (HS256, demo secret "dev-secret")
```
eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.eyJpc3MiOiJodHRwczovL2F1dGguc2hvcGZyb250LmV4YW1wbGUiLCJzdWIiOiJjdXN0Xzg4NDEiLCJhdWQiOlsiYXBpOi8vb3JkZXJzIiwiYXBpOi8vcHJvZmlsZSJdLCJpYXQiOjE3OTIzOTUwMDAsImV4cCI6MTc5Mjk5OTgwMCwiZW1haWwiOiJtaW5hQGV4YW1wbGUuY29tIiwicm9sZXMiOlsiY3VzdG9tZXIiXSwiY2FyZF9sYXN0NCI6IjQyNDIifQ.xupeoN2LYDYe-FKD_4qOh7KCw6G68Zi1pCms7dtUGh0
```
Expected: exit 0; medium lifetime_long (7.0d over 1.0d policy), medium sensitive_claim (card_last4). With `--secret-env` set to "dev-secret": signature verified, plus medium weak_secret (10 bytes).
## 2. Unsigned admin token (alg none)
```
eyJhbGciOiJub25lIiwidHlwIjoiSldUIn0.eyJpc3MiOiJodHRwczovL2F1dGguc2hvcGZyb250LmV4YW1wbGUiLCJzdWIiOiJhZG1pbiIsInJvbGUiOiJhZG1pbiIsImlhdCI6MTc5MjM5NTAwMH0.
```
Expected: exit 1; high alg_none, high aud_mismatch (no aud but api://orders required), medium exp_missing.
## 3. Remote key URL and millisecond expiry (RS256)
```
eyJhbGciOiJSUzI1NiIsInR5cCI6IkpXVCIsImprdSI6Imh0dHBzOi8va2V5cy5hdHRhY2tlci5leGFtcGxlL2p3a3MuanNvbiJ9.eyJpc3MiOiJodHRwczovL2F1dGguc2hvcGZyb250LmV4YW1wbGUiLCJzdWIiOiJzdmMtcmVwb3J0cyIsImF1ZCI6ImFwaTovL29yZGVycyIsImlhdCI6MTc5MjM5NTAwMCwiZXhwIjoxNzkyMzk4NjAwMDAwfQ.bm90LWEtcmVhbC1zaWduYXR1cmUtanVzdC1hLWRlbW8
```
Expected: exit 1; high exp_milliseconds, high header_jku, low kid_missing.
FILE:scripts/jwt_inspect.py
#!/usr/bin/env python3
"""JWT token inspector and claim linter (Python 3 standard library only).
Decodes a JSON Web Token (JWS compact form: header.payload.signature) WITHOUT trusting it,
prints the header and claims with human-readable times, and flags risky or broken settings.
Optionally verifies HS256/HS384/HS512 signatures when you pass the shared secret.
Usage:
python3 jwt_inspect.py <token> # or "-" to read the token from stdin
python3 jwt_inspect.py <token> --now 2026-10-19T09:00:00Z --leeway 60
python3 jwt_inspect.py <token> --expect-iss https://auth.example.com --expect-aud api://orders
python3 jwt_inspect.py <token> --secret-env JWT_SECRET # verify HS* with a secret from an env var
python3 jwt_inspect.py <token> --max-lifetime 3600 --json
python3 jwt_inspect.py --self-test
Severity: high (token unsafe or unusable), medium (likely bug or weak policy), low (hygiene).
Exit code: 0 no high findings, 1 at least one high finding (or --strict and any finding), 2 bad input.
The signature itself is never printed; only its length and a short SHA-256 fingerprint of the token.
"""
import argparse
import base64
import datetime as dt
import hashlib
import hmac
import json
import os
import re
import sys
HS_ALGS = {"HS256": hashlib.sha256, "HS384": hashlib.sha384, "HS512": hashlib.sha512}
MIN_HS_SECRET = {"HS256": 32, "HS384": 48, "HS512": 64}
KNOWN_ALGS = set(HS_ALGS) | {"RS256", "RS384", "RS512", "PS256", "PS384", "PS512",
"ES256", "ES384", "ES512", "EdDSA"}
SENSITIVE = re.compile(r"(pass(word)?|pwd|secret|ssn|social|credit|card|cvv|iban|private|api_?key|refresh_?token)", re.I)
TIME_CLAIMS = ("exp", "nbf", "iat")
class TokenError(ValueError):
pass
def b64url_decode(part, what):
if not re.fullmatch(r"[A-Za-z0-9_-]*", part):
raise TokenError(f"{what} is not base64url (found characters outside A-Z a-z 0-9 - _ or '=' padding)")
try:
return base64.urlsafe_b64decode(part + "=" * (-len(part) % 4))
except (ValueError, TypeError) as e:
raise TokenError(f"{what} could not be base64url-decoded: {e}")
def b64url_encode(raw):
return base64.urlsafe_b64encode(raw).rstrip(b"=").decode("ascii")
def decode(token):
token = token.strip()
if token.lower().startswith("bearer "):
token = token[7:].strip()
parts = token.split(".")
if len(parts) == 5:
raise TokenError("5 segments: this is an encrypted JWE, not a signed JWS; it cannot be inspected without the key")
if len(parts) != 3:
raise TokenError(f"expected 3 dot-separated segments, found {len(parts)}")
out = {}
for name, part in (("header", parts[0]), ("payload", parts[1])):
raw = b64url_decode(part, name)
try:
obj = json.loads(raw.decode("utf-8"))
except (UnicodeDecodeError, ValueError):
raise TokenError(f"{name} is not valid UTF-8 JSON")
if not isinstance(obj, dict):
raise TokenError(f"{name} must be a JSON object")
out[name] = obj
out["signature"] = b64url_decode(parts[2], "signature")
out["signing_input"] = (parts[0] + "." + parts[1]).encode("ascii")
out["fingerprint"] = hashlib.sha256(token.encode("utf-8")).hexdigest()[:12]
return out
def parse_now(value):
if value is None:
return int(dt.datetime.now(dt.timezone.utc).timestamp())
if re.fullmatch(r"\d+", value):
return int(value)
t = dt.datetime.fromisoformat(value.replace("Z", "+00:00"))
if t.tzinfo is None:
t = t.replace(tzinfo=dt.timezone.utc)
return int(t.timestamp())
def iso(ts):
try:
return dt.datetime.fromtimestamp(ts, dt.timezone.utc).strftime("%Y-%m-%dT%H:%M:%SZ")
except (OverflowError, OSError, ValueError, TypeError):
return "invalid"
def human_delta(seconds):
s = abs(int(seconds))
for unit, size in (("d", 86400), ("h", 3600), ("m", 60)):
if s >= size:
return f"{s / size:.1f}{unit}"
return f"{s}s"
def verify_hs(tok, secret):
alg = tok["header"].get("alg")
fn = HS_ALGS.get(alg)
if fn is None:
return None
mac = hmac.new(secret.encode("utf-8"), tok["signing_input"], fn).digest()
return hmac.compare_digest(mac, tok["signature"])
def inspect(token, now=None, leeway=0, expect_iss=None, expect_aud=None, max_lifetime=86400, secret=None):
tok = decode(token)
h, p = tok["header"], tok["payload"]
findings = []
def add(sev, code, msg):
findings.append({"severity": sev, "code": code, "message": msg})
alg = h.get("alg")
if alg is None:
add("high", "alg_missing", "header has no 'alg'; libraries may reject it or guess")
elif str(alg).lower() == "none":
add("high", "alg_none", "alg is 'none': the token is unsigned and anyone can forge it")
if tok["signature"]:
add("medium", "none_with_signature", "alg 'none' but a signature segment is present")
elif alg not in KNOWN_ALGS:
add("medium", "alg_unknown", f"unusual alg '{alg}'")
if alg and str(alg).lower() != "none" and not tok["signature"]:
add("high", "signature_empty", "signature segment is empty although alg requires one")
for key in ("jku", "x5u"):
if key in h:
add("high", f"header_{key}", f"header '{key}' points to a remote key URL ({h[key]}); never fetch keys from a URL inside the token")
if "jwk" in h:
add("high", "header_jwk", "header embeds its own 'jwk'; a verifier that trusts it accepts attacker keys")
if "crit" in h:
add("medium", "header_crit", f"'crit' lists extensions {h['crit']}; verifiers must understand them or reject the token")
if alg in KNOWN_ALGS and alg not in HS_ALGS and "kid" not in h:
add("low", "kid_missing", "asymmetric alg without 'kid'; key rotation will be painful")
now = parse_now(now) if not isinstance(now, int) else now
times = {}
for c in TIME_CLAIMS:
if c in p:
v = p[c]
if isinstance(v, bool) or not isinstance(v, (int, float)):
add("high", f"{c}_not_numeric", f"'{c}' must be NumericDate seconds, got {v!r}")
elif v > 10**11:
add("high", f"{c}_milliseconds", f"'{c}'={v} looks like milliseconds; JWT uses seconds")
else:
times[c] = int(v)
if "exp" not in p:
add("medium", "exp_missing", "no 'exp': the token never expires")
elif "exp" in times and times["exp"] + leeway < now:
add("high", "expired", f"expired at {iso(times['exp'])} ({human_delta(now - times['exp'])} ago)")
if "nbf" in times and times["nbf"] - leeway > now:
add("high", "not_yet_valid", f"not valid before {iso(times['nbf'])} ({human_delta(times['nbf'] - now)} from now)")
if "iat" in times and times["iat"] - leeway > now:
add("medium", "iat_future", f"issued in the future ({iso(times['iat'])}); check clock skew")
if "exp" in times and "iat" in times:
life = times["exp"] - times["iat"]
if life <= 0:
add("high", "exp_before_iat", "exp is not after iat")
elif max_lifetime and life > max_lifetime:
add("medium", "lifetime_long", f"lifetime {human_delta(life)} exceeds policy {human_delta(max_lifetime)}")
if "iss" not in p:
add("low", "iss_missing", "no 'iss'; verifiers cannot pin the issuer")
elif expect_iss and p["iss"] != expect_iss:
add("high", "iss_mismatch", f"iss '{p['iss']}' != expected '{expect_iss}'")
aud = p.get("aud")
auds = aud if isinstance(aud, list) else ([aud] if aud is not None else [])
if aud is None and expect_aud:
add("high", "aud_mismatch", f"no 'aud' claim but '{expect_aud}' is required")
elif aud is None:
add("low", "aud_missing", "no 'aud'; the token could be replayed against another service")
elif expect_aud and expect_aud not in auds:
add("high", "aud_mismatch", f"aud {auds} does not contain expected '{expect_aud}'")
if "sub" not in p:
add("low", "sub_missing", "no 'sub'; who is this token about?")
for k in p:
if SENSITIVE.search(k):
add("medium", "sensitive_claim", f"claim '{k}' looks sensitive; JWT payloads are only encoded, not encrypted")
verified = None
if secret is not None:
if alg not in HS_ALGS:
add("low", "verify_skipped", f"--secret only verifies HS256/384/512; alg is {alg}")
else:
verified = verify_hs(tok, secret)
if not verified:
add("high", "bad_signature", "HMAC signature does NOT match the given secret")
if len(secret.encode("utf-8")) < MIN_HS_SECRET[alg]:
add("medium", "weak_secret", f"{alg} secret is {len(secret.encode('utf-8'))} bytes; use at least {MIN_HS_SECRET[alg]} random bytes")
order = {"high": 0, "medium": 1, "low": 2}
findings.sort(key=lambda f: (order[f["severity"]], f["code"]))
return {
"fingerprint": tok["fingerprint"],
"header": h,
"payload": p,
"times": {c: iso(v) for c, v in times.items()},
"now": iso(now),
"signature_bytes": len(tok["signature"]),
"signature_verified": verified,
"findings": findings,
}
def to_markdown(r):
lines = [f"# JWT inspection (sha256 {r['fingerprint']})", "",
f"Checked at {r['now']}. Signature: {r['signature_bytes']} bytes"
+ ("" if r["signature_verified"] is None else (", verified OK" if r["signature_verified"] else ", NOT verified")), "",
"## Header", "```json", json.dumps(r["header"], indent=2, ensure_ascii=False), "```", "",
"## Claims", "```json", json.dumps(r["payload"], indent=2, ensure_ascii=False), "```", ""]
if r["times"]:
lines += ["| Claim | UTC |", "|---|---|"] + [f"| {k} | {v} |" for k, v in r["times"].items()] + [""]
lines += ["## Findings"]
if not r["findings"]:
lines.append("No findings.")
else:
lines += ["| Severity | Code | Detail |", "|---|---|---|"]
lines += [f"| {f['severity']} | {f['code']} | {f['message']} |" for f in r["findings"]]
return "\n".join(lines)
def make_token(header, payload, secret=None, signature=b""):
si = b64url_encode(json.dumps(header, separators=(",", ":")).encode()) + "." + \
b64url_encode(json.dumps(payload, separators=(",", ":")).encode())
if secret is not None:
signature = hmac.new(secret.encode(), si.encode(), HS_ALGS[header["alg"]]).digest()
return si + "." + b64url_encode(signature)
def self_test():
now = 1760860800 # 2025-10-19T08:00:00Z
strong = "k" * 40
good_claims = {"iss": "https://auth.example.com", "sub": "user-42", "aud": "api://orders",
"iat": now - 60, "exp": now + 840}
good = make_token({"alg": "HS256", "typ": "JWT"}, good_claims, strong)
codes = lambda r: {f["code"] for f in r["findings"]}
cases = []
def check(name, cond):
cases.append((name, bool(cond)))
r = inspect(good, now=now, secret=strong, expect_iss="https://auth.example.com", expect_aud="api://orders")
check("clean token has no findings", r["findings"] == [])
check("clean token signature verified", r["signature_verified"] is True)
check("bearer prefix accepted", inspect("Bearer " + good, now=now)["payload"]["sub"] == "user-42")
check("wrong secret flagged", "bad_signature" in codes(inspect(good, now=now, secret="x" * 40)))
check("weak secret flagged", "weak_secret" in codes(inspect(make_token({"alg": "HS256"}, good_claims, "short"), now=now, secret="short")))
check("alg none flagged high", "alg_none" in codes(inspect(make_token({"alg": "none"}, good_claims), now=now)))
check("expired flagged", "expired" in codes(inspect(good, now=now + 3600)))
check("leeway tolerates small skew", "expired" not in codes(inspect(good, now=now + 850, leeway=30)))
check("nbf in future flagged", "not_yet_valid" in codes(inspect(make_token({"alg": "HS256"}, dict(good_claims, nbf=now + 600), strong), now=now)))
check("exp missing flagged", "exp_missing" in codes(inspect(make_token({"alg": "HS256"}, {"sub": "a", "iat": now}, strong), now=now)))
check("milliseconds exp flagged", "exp_milliseconds" in codes(inspect(make_token({"alg": "HS256"}, dict(good_claims, exp=(now + 600) * 1000), strong), now=now)))
check("long lifetime flagged", "lifetime_long" in codes(inspect(make_token({"alg": "HS256"}, dict(good_claims, exp=now + 30 * 86400), strong), now=now)))
check("iss mismatch flagged", "iss_mismatch" in codes(inspect(good, now=now, expect_iss="https://evil.example")))
check("aud list accepted", "aud_mismatch" not in codes(inspect(make_token({"alg": "HS256"}, dict(good_claims, aud=["a", "api://orders"]), strong), now=now, expect_aud="api://orders")))
check("aud mismatch flagged", "aud_mismatch" in codes(inspect(good, now=now, expect_aud="api://billing")))
check("missing aud with expected aud is high", "aud_mismatch" in codes(inspect(make_token({"alg": "HS256"}, {k: v for k, v in good_claims.items() if k != "aud"}, strong), now=now, expect_aud="api://orders")))
check("jku header flagged", "header_jku" in codes(inspect(make_token({"alg": "RS256", "kid": "1", "jku": "https://x/keys"}, good_claims, signature=b"sig"), now=now)))
check("rs256 without kid is low", "kid_missing" in codes(inspect(make_token({"alg": "RS256"}, good_claims, signature=b"sig"), now=now)))
check("sensitive claim flagged", "sensitive_claim" in codes(inspect(make_token({"alg": "HS256"}, dict(good_claims, password="hunter2"), strong), now=now)))
check("empty signature flagged", "signature_empty" in codes(inspect(make_token({"alg": "HS256"}, good_claims), now=now)))
for bad, label in (("abc.def", "two segments rejected"), ("a.b.c.d.e", "JWE rejected"), ("!!.e30.", "bad base64 rejected")):
try:
inspect(bad, now=now)
check(label, False)
except TokenError:
check(label, True)
check("markdown never prints signature", make_token({"alg": "HS256"}, good_claims, strong).split(".")[2] not in to_markdown(r))
passed = sum(ok for _, ok in cases)
for name, ok in cases:
print(("PASS " if ok else "FAIL ") + name)
print(f"{passed}/{len(cases)} passed")
return 0 if passed == len(cases) else 1
def main(argv=None):
ap = argparse.ArgumentParser(description="Decode and lint a JWT without trusting it.")
ap.add_argument("token", nargs="?", help="JWT string, or '-' to read from stdin")
ap.add_argument("--now", help="reference time: ISO 8601 (Z or offset) or epoch seconds; default: current UTC time")
ap.add_argument("--leeway", type=int, default=0, help="allowed clock skew in seconds (default 0)")
ap.add_argument("--expect-iss")
ap.add_argument("--expect-aud")
ap.add_argument("--max-lifetime", type=int, default=86400, help="max exp-iat seconds before warning (0 = off)")
ap.add_argument("--secret-env", help="name of an environment variable holding the HS* secret (avoids shell history)")
ap.add_argument("--json", action="store_true")
ap.add_argument("--strict", action="store_true", help="exit 1 on any finding")
ap.add_argument("--self-test", action="store_true")
a = ap.parse_args(argv)
if a.self_test:
return self_test()
if not a.token:
ap.error("token required (or --self-test)")
token = sys.stdin.read() if a.token == "-" else a.token
secret = None
if a.secret_env:
secret = os.environ.get(a.secret_env)
if secret is None:
print(f"error: environment variable {a.secret_env} is not set", file=sys.stderr)
return 2
try:
r = inspect(token, now=a.now, leeway=a.leeway, expect_iss=a.expect_iss, expect_aud=a.expect_aud,
max_lifetime=a.max_lifetime, secret=secret)
except (TokenError, ValueError) as e:
print(f"error: {e}", file=sys.stderr)
return 2
print(json.dumps(r, indent=2, ensure_ascii=False) if a.json else to_markdown(r))
if any(f["severity"] == "high" for f in r["findings"]) or (a.strict and r["findings"]):
return 1
return 0
if __name__ == "__main__":
sys.exit(main())Checks CSV time series (metrics, sensor readings, sales by hour) for data problems before you chart or alert on them: missing intervals, duplicate or out-of-order timestamps, empty values, flatlines from stuck exporters, robust median/MAD spikes, and counter resets. Use when a dashboard looks wrong, before training a forecast, or when asking "is this data complete?". Includes a stdlib Python checker with --self-test.
---
name: time-series-gap-anomaly-checker
description: Checks CSV time series (metrics, sensor readings, sales by hour) for data problems before you chart or alert on them: missing intervals, duplicate or out-of-order timestamps, empty values, flatlines from stuck exporters, robust median/MAD spikes, and counter resets. Use when a dashboard looks wrong, before training a forecast, or when asking "is this data complete?". Includes a stdlib Python checker with --self-test.
---
# Time Series Gap and Anomaly Checker
You help analysts, SREs, and data engineers decide whether a time series can be trusted. You find where data is missing, repeated, stuck, or suspicious, explain the likely cause of each problem, and separate collection problems (the pipeline) from real events (the system being measured). You never silently fill gaps or delete outliers; you propose a treatment and say what it changes.
## Files in this skill
- `scripts/ts_check.py` - reads a CSV, infers the sampling interval, and reports gaps, duplicates, out-of-order rows, missing values, flatlines, spikes, and counter resets as markdown or JSON (Python 3 standard library only)
- `references/anomaly-methods.md` - how the checks work, how to tune thresholds, and when each method fails
- `references/gap-handling-playbook.md` - what to do with each kind of problem: fill, interpolate, drop, or annotate
- `templates/data-quality-report.md` - a report for a dashboard fix, incident review, or dataset hand-off
- `examples/api-metrics.csv` - three hours of 5-minute API metrics with a scrape outage, a duplicate, a stuck value, a latency spike, and a counter reset
## Workflow
### 1. Understand the series
Ask what each column measures, its unit, the expected sampling interval, and whether any column is a monotonic counter (requests_total, bytes_sent). Ask which time zone naive timestamps use; the script treats them as UTC.
### 2. Run the checker
```
python3 scripts/ts_check.py examples/api-metrics.csv --counter requests_total
python3 scripts/ts_check.py data.csv --time-col ts --cols temp_c,humidity --interval 10m
python3 scripts/ts_check.py data.csv --z 5 --window 12 --flatline-points 12 --json > report.json
python3 scripts/ts_check.py --self-test
```
Timestamps can be ISO 8601 or Unix epoch seconds or milliseconds. Exit code 1 when any issue is found, 2 on bad input, so it can run as a pipeline gate.
### 3. Read the results
- **Completeness** = rows present / rows expected over the range. Below 95% usually means aggregates and averages are biased.
- **gap**: spacing over 1.5x the interval. Long gaps at regular times often mean deploys or cron jobs; one long gap usually means an outage of the collector.
- **duplicate / out_of_order**: double scrapes, retries, or merged files.
- **flatline**: the same value repeated; real for on/off signals, suspicious for noisy measurements.
- **spike**: robust z-score against a rolling median. Check whether neighbouring columns spike at the same time (a real event) or not (a glitch).
- **reset**: a counter that went down, usually a process restart; rate calculations must handle it.
### 4. Recommend treatment
Use `references/gap-handling-playbook.md`. Tune thresholds with `references/anomaly-methods.md` if the data is seasonal, bursty, or very sparse.
### 5. Report
Fill `templates/data-quality-report.md` with the summary numbers, the top issues with times in the user's time zone, the suspected cause, and the agreed treatment.
## Rules
- Report problems; do not alter the source data. Write cleaned data to a new file and document every change.
- Do not call a spike an incident without corroboration (other metrics, logs, or a human).
- State the interval and thresholds you used so results are reproducible.
- Convert UTC times to the user's time zone when presenting them.
FILE:references/anomaly-methods.md
# Anomaly Methods and Tuning
## Interval inference
The expected interval is the median spacing between sorted unique timestamps. This survives a few gaps and jitter. Pass `--interval` when the series is mostly gaps (sparse events) or when you know the scrape interval.
A gap is any spacing above 1.5x the interval, so 5-minute data with a 7-minute spacing is not a gap but a 10-minute spacing is (1 missing point).
## Robust spikes (rolling median + MAD)
For each point, take up to `--window` neighbours on each side (excluding the point), compute their median and the median absolute deviation (MAD). The robust z-score is:
z = |value - median| / (1.4826 * MAD)
1.4826 makes MAD comparable to a standard deviation for normal data. Default threshold 4 is roughly a 1-in-15,000 event for normal noise. When more than half the neighbours are identical (MAD = 0), the mean absolute deviation is used, and for a perfectly flat neighbourhood 1% of the level.
Why not mean and standard deviation? One large spike inflates the standard deviation and hides itself. The median and MAD ignore it.
## Tuning
| Data | Suggestion |
|---|---|
| Noisy metrics (latency, CPU) | `--z 5`, `--window 12` |
| Daily or weekly seasonality | Check one season at a time, or compare to the same hour last week before trusting spikes |
| Bursty business data (orders) | Higher `--z` (6-8); spikes are often real |
| Integer or low-resolution values | Expect MAD = 0; raise `--z` or aggregate first |
| On/off or status signals | Ignore flatline; check only gaps |
| Counters | Always pass `--counter`; check rates after handling resets |
## Known limits
- Gradual drifts and level shifts are not spikes; plot the series or compare week-over-week medians.
- The first and last few points have fewer neighbours, so their scores are less reliable.
- Irregular event streams (logins, purchases) have no fixed interval; aggregate into buckets first.
FILE:references/gap-handling-playbook.md
# Gap and Anomaly Handling Playbook
| Problem | Likely causes | Default treatment | Avoid |
|---|---|---|---|
| Short gap (1-2 points) | scrape timeout, network blip | Linear interpolation for smooth signals; leave null for counts | Filling counts with zero |
| Long gap | collector down, deploy, host off | Leave null and annotate the chart; exclude from SLO math or count as unknown | Interpolating across hours |
| Duplicate timestamps | double scrape, retried batch, merged files | Keep one row (first, or the one with most fields); check whether values differ | Summing duplicates |
| Out-of-order rows | late-arriving data, file concatenation | Sort by time before any windowed calculation | Assuming file order is time order |
| Missing values | sensor dropout, parse errors, "N/A" strings | Fix the parser first; then treat as a gap | Treating blank as 0 |
| Flatline | stuck exporter, cached value, sensor freeze | Confirm with a second source; mark as suspect | Trusting it for alerts |
| Spike (single point) | glitch, unit error, real burst | Check related columns and logs; cap or remove only with a documented reason | Deleting without a note |
| Counter reset | process restart, overflow | Compute rates per segment (rate after reset = new value / interval) | Negative rates |
## Before modeling or forecasting
1. Sort, de-duplicate, and regularize to the expected interval (insert nulls for missing slots).
2. Decide gap policy per column and record it.
3. Keep a mask column (`is_imputed`) so models and readers know what was filled.
4. Re-run the checker on the cleaned file and attach both reports.
## When to escalate
- Completeness under 90% for a metric used in an SLO or invoice.
- Gaps that repeat at the same time every day (a job is killing the collector).
- Spikes that appear in all hosts at once (a real incident, not a data issue).
FILE:templates/data-quality-report.md
# Time Series Data Quality - <dataset or dashboard>
**Source:** <file / query> **Range:** <start> to <end> (<time zone>) **Interval:** <expected>
**Rows:** <n> **Completeness:** <percent> **Thresholds:** z=<z>, window=<w>, flatline=<n>
## Summary
<two sentences: can this data be trusted for the intended use, and what must change first>
## Issues
| Kind | When (<time zone>) | Column | Detail | Suspected cause | Treatment |
|---|---|---|---|---|---|
| | | | | | |
## Decisions
- Gap policy: <interpolate short gaps / leave null / exclude>
- Outlier policy: <keep / cap / remove with note>
- Counter handling: <segment rates at resets>
## Follow-ups
- [ ] <fix collector / parser / alert> - owner <name> - due <date>
Generated with scripts/ts_check.py. Source data was not modified.
FILE:examples/api-metrics.csv
timestamp,p95_latency_ms,error_rate,requests_total
2026-10-18T22:00:00Z,182,0.4,1200
2026-10-18T22:05:00Z,176,0.3,2435
2026-10-18T22:10:00Z,190,0.5,3705
2026-10-18T22:15:00Z,185,0.4,5010
2026-10-18T22:20:00Z,179,0.3,6210
2026-10-18T22:25:00Z,188,0.4,7445
2026-10-18T22:30:00Z,181,0.5,8715
2026-10-18T22:35:00Z,193,0.4,10020
2026-10-18T22:35:00Z,193,0.4,10020
2026-10-18T22:40:00Z,177,0.3,11220
2026-10-18T22:45:00Z,184,0.4,12455
2026-10-18T22:50:00Z,186,0.4,13725
2026-10-18T22:55:00Z,180,0.4,15030
2026-10-18T23:00:00Z,191,0.4,16230
2026-10-18T23:05:00Z,,0.4,17465
2026-10-18T23:10:00Z,178,0.4,18735
2026-10-18T23:15:00Z,187,0.4,20040
2026-10-18T23:20:00Z,182,0.4,21240
2026-10-18T23:25:00Z,189,0.4,22475
2026-10-18T23:30:00Z,612,2.1,23745
2026-10-18T23:35:00Z,185,0.5,25050
2026-10-18T23:40:00Z,180,0.3,26250
2026-10-18T23:45:00Z,184,0.4,27485
2026-10-19T00:10:00Z,188,0.5,28755
2026-10-19T00:15:00Z,183,0.4,30060
2026-10-19T00:20:00Z,177,0.3,31260
2026-10-19T00:25:00Z,185,0.4,32495
2026-10-19T00:30:00Z,182,0.5,1180
2026-10-19T00:35:00Z,189,0.4,2485
2026-10-19T00:40:00Z,184,0.3,3685
2026-10-19T00:45:00Z,180,0.4,4920
2026-10-19T00:50:00Z,186,0.5,6190
2026-10-19T00:55:00Z,183,0.4,7495
FILE:scripts/ts_check.py
#!/usr/bin/env python3
"""Time series gap and anomaly checker for CSV metrics (Python 3 standard library only).
Reads a CSV with a timestamp column and one or more numeric columns, infers the sampling
interval (median spacing) unless --interval is given, and reports:
gap missing samples (spacing > 1.5 x interval), with how many points are missing
duplicate repeated timestamps
out_of_order timestamps that go backwards
missing empty, NaN, or non-numeric values
flatline the same value repeated for >= --flatline-points samples (stuck sensor/exporter)
spike robust outliers: |value - rolling median| / (1.4826 * MAD) >= --z (default 4);
if MAD is 0 the mean absolute deviation is used instead
reset a counter column that decreases (--counter col1,col2)
Timestamps: ISO 8601 (with Z or offset; naive = UTC) or Unix epoch seconds/milliseconds.
Usage:
python3 ts_check.py metrics.csv
python3 ts_check.py metrics.csv --time-col ts --cols cpu,latency_ms --interval 5m --z 5
python3 ts_check.py metrics.csv --counter requests_total --json
python3 ts_check.py --self-test
Exit code: 0 clean, 1 issues found, 2 bad input.
"""
import argparse
import csv
import datetime as dt
import io
import json
import math
import re
import statistics
import sys
class InputError(ValueError):
pass
def parse_ts(raw):
s = (raw or "").strip()
if not s:
raise InputError("empty timestamp")
if re.fullmatch(r"-?\d+(\.\d+)?", s):
v = float(s)
return v / 1000.0 if v > 1e11 else v
try:
t = dt.datetime.fromisoformat(s.replace("Z", "+00:00").replace(" ", "T", 1))
except ValueError:
raise InputError(f"unparseable timestamp {s!r}")
if t.tzinfo is None:
t = t.replace(tzinfo=dt.timezone.utc)
return t.timestamp()
def parse_interval(s):
m = re.fullmatch(r"(\d+(?:\.\d+)?)\s*(ms|s|m|h|d)?", s.strip())
if not m:
raise InputError(f"bad interval {s!r}; use e.g. 30s, 5m, 1h")
mult = {"ms": 0.001, "s": 1, None: 1, "m": 60, "h": 3600, "d": 86400}[m.group(2)]
return float(m.group(1)) * mult
def fmt_ts(ts):
return dt.datetime.fromtimestamp(ts, dt.timezone.utc).strftime("%Y-%m-%dT%H:%M:%SZ")
def fmt_dur(sec):
for unit, size in (("d", 86400), ("h", 3600), ("m", 60)):
if sec >= size:
v = sec / size
return f"{v:.0f}{unit}" if v == int(v) else f"{v:.1f}{unit}"
return f"{sec:g}s"
def to_float(raw):
s = (raw or "").strip()
if s == "" or s.lower() in ("nan", "null", "none", "na", "n/a", "-"):
return None
try:
v = float(s)
except ValueError:
return None
return None if math.isnan(v) or math.isinf(v) else v
def load(text, time_col=None, cols=None):
reader = csv.DictReader(io.StringIO(text))
if not reader.fieldnames:
raise InputError("CSV has no header row")
fields = [f.strip() for f in reader.fieldnames]
reader.fieldnames = fields
if time_col is None:
cands = [f for f in fields if f.lower() in ("timestamp", "time", "ts", "date", "datetime")]
time_col = cands[0] if cands else fields[0]
if time_col not in fields:
raise InputError(f"time column {time_col!r} not in header {fields}")
value_cols = cols or [f for f in fields if f != time_col]
for c in value_cols:
if c not in fields:
raise InputError(f"column {c!r} not in header {fields}")
rows = []
for i, r in enumerate(reader, start=2):
rows.append({"line": i, "ts": parse_ts(r.get(time_col)),
"raw": {c: r.get(c) for c in value_cols}})
if len(rows) < 3:
raise InputError("need at least 3 data rows")
return time_col, value_cols, rows
def rolling_robust(values, idx, window):
lo, hi = max(0, idx - window), min(len(values), idx + window + 1)
neigh = [values[j] for j in range(lo, hi) if j != idx and values[j] is not None]
if len(neigh) < 4:
return None, None
med = statistics.median(neigh)
mad = statistics.median(abs(v - med) for v in neigh)
scale = 1.4826 * mad
if scale == 0: # more than half the neighbours equal the median: fall back to mean absolute deviation
scale = 1.2533 * statistics.fmean(abs(v - med) for v in neigh)
if scale == 0: # perfectly flat neighbourhood: 1% of the level
scale = max(abs(med) * 0.01, 1e-9)
return med, scale
def check(text, time_col=None, cols=None, interval=None, z=4.0, window=6, flatline_points=6, counters=()):
time_col, value_cols, rows = load(text, time_col, cols)
issues = []
def add(kind, line, ts, col, detail):
issues.append({"kind": kind, "line": line, "time": fmt_ts(ts), "column": col, "detail": detail})
# order and duplicates (in file order)
seen = {}
for prev, cur in zip(rows, rows[1:]):
if cur["ts"] < prev["ts"]:
add("out_of_order", cur["line"], cur["ts"], time_col, f"goes back {fmt_dur(prev['ts'] - cur['ts'])} from line {prev['line']}")
for r in rows:
if r["ts"] in seen:
add("duplicate", r["line"], r["ts"], time_col, f"same timestamp as line {seen[r['ts']]}")
else:
seen[r["ts"]] = r["line"]
srt = sorted({r["ts"]: r for r in reversed(rows)}.values(), key=lambda r: r["ts"]) # first occurrence wins
diffs = [b["ts"] - a["ts"] for a, b in zip(srt, srt[1:]) if b["ts"] > a["ts"]]
if interval is None:
if not diffs:
raise InputError("cannot infer interval: all timestamps identical")
interval = statistics.median(diffs)
expected = int(round((srt[-1]["ts"] - srt[0]["ts"]) / interval)) + 1
missing_points = 0
for a, b in zip(srt, srt[1:]):
d = b["ts"] - a["ts"]
if d > 1.5 * interval:
n = int(round(d / interval)) - 1
missing_points += n
add("gap", b["line"], a["ts"], time_col, f"{fmt_dur(d)} without data after {fmt_ts(a['ts'])} (~{n} missing)")
stats = {}
for c in value_cols:
vals = [to_float(r["raw"][c]) for r in srt]
for r, v in zip(srt, vals):
if v is None:
add("missing", r["line"], r["ts"], c, f"value {r['raw'][c]!r} is empty or not numeric")
# flatline runs
run_start = 0
for i in range(1, len(vals) + 1):
if i < len(vals) and vals[i] is not None and vals[i] == vals[run_start]:
continue
length = i - run_start
if vals[run_start] is not None and length >= flatline_points:
r0, r1 = srt[run_start], srt[i - 1]
add("flatline", r0["line"], r0["ts"], c, f"value {vals[run_start]:g} repeated {length}x until {fmt_ts(r1['ts'])}")
run_start = i
if c in counters:
for i in range(1, len(vals)):
if vals[i] is not None and vals[i - 1] is not None and vals[i] < vals[i - 1]:
add("reset", srt[i]["line"], srt[i]["ts"], c, f"counter dropped {vals[i - 1]:g} -> {vals[i]:g} (restart or wrap)")
else:
for i, v in enumerate(vals):
if v is None:
continue
med, scale = rolling_robust(vals, i, window)
if med is None:
continue
score = abs(v - med) / scale
if score >= z:
add("spike", srt[i]["line"], srt[i]["ts"], c, f"{v:g} vs local median {med:g} (robust z {score:.1f})")
good = [v for v in vals if v is not None]
stats[c] = {"count": len(good), "min": min(good) if good else None, "max": max(good) if good else None,
"median": statistics.median(good) if good else None}
order = {"out_of_order": 0, "duplicate": 1, "gap": 2, "reset": 3, "missing": 4, "flatline": 5, "spike": 6}
issues.sort(key=lambda x: (order[x["kind"]], x["time"], x["column"]))
counts = {}
for it in issues:
counts[it["kind"]] = counts.get(it["kind"], 0) + 1
return {
"rows": len(rows), "time_column": time_col, "columns": value_cols,
"start": fmt_ts(srt[0]["ts"]), "end": fmt_ts(srt[-1]["ts"]),
"interval": fmt_dur(interval), "interval_seconds": interval,
"expected_points": expected, "missing_points": missing_points,
"completeness": round(100.0 * len(srt) / expected, 1) if expected else 100.0,
"stats": stats, "counts": counts, "issues": issues,
}
def to_markdown(r):
lines = ["# Time series check", "",
f"Rows: {r['rows']} Range: {r['start']} to {r['end']} Interval: {r['interval']}",
f"Completeness: {r['completeness']}% ({r['missing_points']} of {r['expected_points']} expected points missing)", "",
"| Column | Count | Min | Median | Max |", "|---|---|---|---|---|"]
for c, s in r["stats"].items():
f = lambda v: "-" if v is None else f"{v:g}"
lines.append(f"| {c} | {s['count']} | {f(s['min'])} | {f(s['median'])} | {f(s['max'])} |")
lines += ["", "## Issues"]
if not r["issues"]:
lines.append("No issues found.")
else:
lines.append("Totals: " + ", ".join(f"{k} {v}" for k, v in r["counts"].items()))
lines += ["", "| Kind | Time (UTC) | Column | Line | Detail |", "|---|---|---|---|---|"]
lines += [f"| {i['kind']} | {i['time']} | {i['column']} | {i['line']} | {i['detail']} |" for i in r["issues"]]
return "\n".join(lines)
def self_test():
cases = []
def ok(name, cond):
cases.append((name, bool(cond)))
base = 1760860800 # 2025-10-19T08:00:00Z
def csv_of(points, header="timestamp,value"):
return header + "\n" + "\n".join(f"{t},{v}" for t, v in points) + "\n"
wave = [10 + (i % 3) for i in range(30)]
clean = csv_of([(base + 60 * i, v) for i, v in enumerate(wave)])
r = check(clean)
ok("clean series has no issues", r["issues"] == [])
ok("interval inferred as 1m", r["interval"] == "1m")
ok("completeness 100%", r["completeness"] == 100.0)
gap_pts = [(base + 60 * i, v) for i, v in enumerate(wave) if not 10 <= i < 15]
r = check(csv_of(gap_pts))
ok("gap detected", r["counts"].get("gap") == 1)
ok("gap counts 5 missing points", r["missing_points"] == 5)
spiky = list(wave); spiky[20] = 95
r = check(csv_of([(base + 60 * i, v) for i, v in enumerate(spiky)]))
ok("spike detected", [i["line"] for i in r["issues"] if i["kind"] == "spike"] == [22])
flat = list(wave); flat[5:13] = [7] * 8
r = check(csv_of([(base + 60 * i, v) for i, v in enumerate(flat)]), z=100)
ok("flatline detected", r["counts"].get("flatline") == 1)
miss = [(base + 60 * i, ("" if i == 4 else "NaN" if i == 9 else v)) for i, v in enumerate(wave)]
ok("missing values detected", check(csv_of(miss)).get("counts").get("missing") == 2)
dup = [(base + 60 * i, v) for i, v in enumerate(wave)]
dup.insert(6, dup[5])
ok("duplicate detected", check(csv_of(dup))["counts"].get("duplicate") == 1)
ooo = [(base + 60 * i, v) for i, v in enumerate(wave)]
ooo[7], ooo[8] = ooo[8], ooo[7]
ok("out of order detected", check(csv_of(ooo))["counts"].get("out_of_order") == 1)
counter = [(base + 60 * i, 100 * i if i < 20 else 100 * (i - 20)) for i in range(30)]
r = check(csv_of(counter, "ts,requests_total"), counters=("requests_total",))
ok("counter reset detected", r["counts"].get("reset") == 1 and "spike" not in r["counts"])
iso_rows = "time,cpu\n" + "\n".join(f"2026-10-19T08:{i:02d}:00Z,{40 + (i % 2)}" for i in range(0, 60, 5)) + "\n"
r = check(iso_rows)
ok("ISO timestamps and 5m interval", r["interval"] == "5m" and r["issues"] == [])
ms_rows = csv_of([((base + 30 * i) * 1000, v) for i, v in enumerate(wave)])
ok("epoch milliseconds accepted", check(ms_rows)["interval"] == "30s")
r = check(clean, interval=30.0)
ok("explicit 30s interval flags every 1m step as a gap", r["counts"].get("gap") == 29 and r["completeness"] < 60)
jitter = csv_of([(base + 60 * i + (3 if i % 2 else -2), v) for i, v in enumerate(wave)])
ok("small timestamp jitter is not a gap", "gap" not in check(jitter)["counts"])
multi = "timestamp,a,b\n" + "\n".join(f"{base + 60 * i},{wave[i]},{wave[i] * 2 if i != 12 else 500}" for i in range(30)) + "\n"
r = check(multi, cols=["a", "b"])
ok("multi-column spike only in b", {i["column"] for i in r["issues"]} == {"b"})
for bad, label in (("timestamp,value\n1,2\n", "too few rows rejected"),
("timestamp,value\nyesterday,1\ntoday,2\nnow,3\n", "bad timestamp rejected")):
try:
check(bad)
ok(label, False)
except InputError:
ok(label, True)
ok("markdown renders", "## Issues" in to_markdown(check(clean)))
passed = sum(c for _, c in cases)
for name, c in cases:
print(("PASS " if c else "FAIL ") + name)
print(f"{passed}/{len(cases)} passed")
return 0 if passed == len(cases) else 1
def main(argv=None):
ap = argparse.ArgumentParser(description="Find gaps, duplicates, flatlines, spikes, and counter resets in CSV metrics.")
ap.add_argument("csv", nargs="?", help="CSV file, or '-' for stdin")
ap.add_argument("--time-col")
ap.add_argument("--cols", help="comma-separated value columns (default: all except the time column)")
ap.add_argument("--interval", help="expected spacing, e.g. 30s, 5m, 1h (default: median spacing)")
ap.add_argument("--z", type=float, default=4.0, help="robust z-score threshold for spikes (default 4)")
ap.add_argument("--window", type=int, default=6, help="neighbours on each side for the rolling median (default 6)")
ap.add_argument("--flatline-points", type=int, default=6)
ap.add_argument("--counter", default="", help="comma-separated monotonic counter columns (checked for resets, not spikes)")
ap.add_argument("--json", action="store_true")
ap.add_argument("--self-test", action="store_true")
a = ap.parse_args(argv)
if a.self_test:
return self_test()
if not a.csv:
ap.error("csv path required (or --self-test)")
try:
text = sys.stdin.read() if a.csv == "-" else open(a.csv, encoding="utf-8-sig").read()
r = check(text, a.time_col, [c.strip() for c in a.cols.split(",")] if a.cols else None,
parse_interval(a.interval) if a.interval else None, a.z, a.window, a.flatline_points,
tuple(c.strip() for c in a.counter.split(",") if c.strip()))
except (OSError, InputError) as e:
print(f"error: {e}", file=sys.stderr)
return 2
print(json.dumps(r, indent=2) if a.json else to_markdown(r))
return 1 if r["issues"] else 0
if __name__ == "__main__":
sys.exit(main())Splits shared costs for a group trip, house share, or event: equal, subset, weighted, or exact splits, mixed currencies with stated rates, and repayments already made, all in integer cents so totals match. Produces per-person balances and a short "who pays whom" list (at most one transfer fewer than people). Use when friends ask "who owes what?" after a trip or a shared house month. Includes a stdlib Python splitter with --self-test.
---
name: group-trip-expense-splitter
description: Splits shared costs for a group trip, house share, or event: equal, subset, weighted, or exact splits, mixed currencies with stated rates, and repayments already made, all in integer cents so totals match. Produces per-person balances and a short "who pays whom" list (at most one transfer fewer than people). Use when friends ask "who owes what?" after a trip or a shared house month. Includes a stdlib Python splitter with --self-test.
---
# Group Trip Expense Splitter
You help a group settle shared expenses fairly and without arguments. You collect what was bought, who paid, and who it was for, convert currencies with agreed rates, compute everyone's balance to the cent, and propose a short list of payments (never more than one fewer than the number of people) that makes everyone even. You explain every number so anyone in the group can check it.
## Files in this skill
- `scripts/split_expenses.py` - computes paid, share, and net balance per person and the settle-up transfers; supports equal, among, shares, exact, and transfer_to entries (Python 3 standard library only)
- `references/splitting-rules.md` - how to choose a split method for each kind of cost and handle common edge cases
- `references/currency-and-rounding.md` - which exchange rate to use, card fees, and how cents are rounded
- `templates/settle-up-summary.md` - a summary to paste into the group chat
- `examples/trip.json` - a 4-person weekend with a skipped dinner, student tickets, weighted drinks, USD and GBP costs, and a cash repayment
## Workflow
### 1. Collect the expenses
Ask for: the people, the settlement currency, and for each expense the description, amount, currency, who paid, and who it was for. Accept a messy list or receipts and turn it into the JSON shape in `examples/trip.json`. Ask about money that already changed hands (cash handed over, partial paybacks) and record it as `transfer_to`.
### 2. Choose the split for each item
Default is equal among everyone. Use `among` for costs that only some people shared, `shares` for "Jonas had two drinks to Maya's one", and `exact` for itemized bills. See `references/splitting-rules.md` and confirm unclear items with the group instead of guessing.
### 3. Set currency rates
Agree on one rate per currency (card statement rate, or a published rate on a stated date) and put it in `rates` as "1 unit = X settlement units". See `references/currency-and-rounding.md`.
### 4. Run the splitter
```
python3 scripts/split_expenses.py examples/trip.json
python3 scripts/split_expenses.py trip.json --json > settle.json
python3 scripts/split_expenses.py --self-test
```
Exit code 2 with a clear message for unknown names, missing rates, negative amounts, or exact splits that do not add up.
### 5. Share the result
Fill `templates/settle-up-summary.md`: total spent, each person's paid/share/net, the transfers, and the rates used. Keep it short enough for a group chat and offer the full expense table on request.
## Rules
- Never guess who shared an expense; ask, or split equally and say so.
- Totals must match to the cent; mention who received rounding cents.
- Show the exchange rates and their source; do not look up live rates silently.
- Stay neutral: present numbers, not judgments about who spent more.
- Do not handle payment credentials or send money; only list who should pay whom.
FILE:references/splitting-rules.md
# Splitting Rules and Edge Cases
## Choosing a method
| Situation | Method | Example |
|---|---|---|
| Everyone used it | equal (default) | accommodation, car rental, shared groceries |
| Only some people took part | `among` | dinner one person skipped, a tour two people joined |
| Uneven use of the same thing | `shares` | drinks, nights stayed (3 nights vs 2), fuel by distance |
| Itemized bill or different prices | `exact` | student vs adult tickets, separate menu items |
| Money already handed over | `transfer_to` | cash paid back, a bank transfer during the trip |
## Edge cases
- **Someone arrives late or leaves early.** Use `shares` by nights for the accommodation (e.g. 3, 3, 2, 3).
- **Couples or families paying as one.** Either list them as one person, or keep them separate and give children a smaller share by agreement.
- **Deposits and refunds.** Amounts must be positive, so lower the original expense by the refund. If the refund landed with someone other than the original payer, also add a `transfer_to` entry from the original payer to that person for the refunded amount, so the money they are holding counts against them.
- **Tips and service charges.** Split like the bill they belong to.
- **Gifts for one member (birthday dinner).** Use `among` with everyone except the guest.
- **Shared kitty / common pot.** Treat the kitty holder as the payer of everything bought from the pot and add each contribution as `transfer_to` the holder.
- **Personal items on a shared receipt.** Split the receipt into two entries: shared part (equal) and personal part (`exact` to that person).
## Fairness checks before sharing
1. Sum of everyone's share equals total spent.
2. Every expense has a payer who was actually there.
3. Nobody's share includes things they explicitly declined.
4. The group agreed on the rates and on any unequal splits.
FILE:references/currency-and-rounding.md
# Currency and Rounding
## Picking exchange rates
- **Best:** the rate on the payer's card or bank statement for that charge, including any foreign transaction fee, because that is what they actually paid.
- **Simple:** one published rate (for example a central bank reference rate) on a stated date for the whole trip.
- **Cash bought at an exchange office:** the rate they bought the cash at.
Write rates as "1 unit of foreign currency = X units of settlement currency": `"USD": 0.92` means 1 USD = 0.92 EUR.
## Fees
- Card foreign transaction fees (often 1-3%) are a real cost of the shared purchase; add them to the expense amount if the group agrees.
- Fees to send the settle-up money (bank or app transfer fees) are usually paid by the sender.
## Rounding to the cent
1. Each expense is converted and rounded once (half up) to whole cents.
2. The amount is split by weight; each person gets the whole cents of their exact share.
3. Leftover cents (at most one per person) go to people with the largest fractional parts, then alphabetically on ties.
This keeps every expense exact to the cent and the result reproducible. Over a trip the rounding difference per person is a few cents at most.
## Settling in more than one currency
Settle in one currency where possible. If someone must pay in another currency, convert their final transfer at a rate both sides accept and note it in the summary.
FILE:templates/settle-up-summary.md
# Settle-up - <trip or event name>
**Dates:** <dates> **People:** <names> **Settlement currency:** <currency>
**Total spent:** <amount> across <n> expenses **Rates used:** <1 USD = 0.92 EUR (card statements), ...>
## Balances
| Person | Paid | Share | Net (+ gets back / - owes) |
|---|---|---|---|
| | | | |
## Who pays whom
- <name> pays <name> <amount>
- <name> pays <name> <amount>
## Notes
- Unequal splits: <dinner without Leo, student tickets, drinks by count>
- Rounding cents went to: <names>
- Already settled during the trip: <cash or transfers>
## Group chat message
> Trip totals are in! We spent <total> together. To settle up: <transfer 1>; <transfer 2>; <transfer 3>. Full breakdown is in the thread. Shout if anything looks off by <date>.
FILE:examples/trip.json
{
"trip": "Lisbon long weekend, 4 friends",
"currency": "EUR",
"rates": {"USD": 0.92, "GBP": 1.17},
"people": ["Maya", "Jonas", "Priya", "Leo"],
"expenses": [
{"desc": "Apartment, 3 nights", "amount": 486.00, "paid_by": "Maya"},
{"desc": "Airport taxi", "amount": 38.50, "paid_by": "Jonas"},
{"desc": "Groceries for breakfasts", "amount": 61.27, "paid_by": "Priya"},
{"desc": "Seafood dinner (Leo skipped)", "amount": 142.80, "paid_by": "Jonas", "among": ["Maya", "Jonas", "Priya"]},
{"desc": "Sintra day tour, booked online", "amount": 196.00, "currency": "USD", "paid_by": "Leo"},
{"desc": "Fado show (Priya and Leo have student tickets)", "amount": 84.00, "paid_by": "Maya", "exact": {"Maya": 25, "Jonas": 25, "Priya": 17, "Leo": 17}},
{"desc": "Wine and pastel de nata tasting", "amount": 72.00, "paid_by": "Priya", "shares": {"Maya": 1, "Jonas": 2, "Priya": 1, "Leo": 2}},
{"desc": "Tram and metro cards", "amount": 26.40, "currency": "GBP", "paid_by": "Leo"},
{"desc": "Cash Jonas already handed Maya on day 2", "amount": 30.00, "paid_by": "Jonas", "transfer_to": "Maya"}
]
}
FILE:scripts/split_expenses.py
#!/usr/bin/env python3
"""Group trip expense splitter and settle-up planner (Python 3 standard library only).
Input JSON:
{
"currency": "EUR", # settlement currency
"rates": {"USD": 0.92, "TRY": 0.027}, # 1 unit of key = rate units of settlement currency
"people": ["Ana", "Ben", "Cem", "Dee"],
"expenses": [
{"desc": "Villa", "amount": 1200, "paid_by": "Ana"}, # equal, everyone
{"desc": "Dinner", "amount": 180, "paid_by": "Ben", "among": ["Ana", "Ben", "Cem"]}, # equal, subset
{"desc": "Boat", "amount": 300, "currency": "USD", "paid_by": "Cem",
"shares": {"Ana": 2, "Ben": 1, "Cem": 1}}, # weighted
{"desc": "Museum", "amount": 54, "paid_by": "Dee", "exact": {"Dee": 18, "Ana": 36}}, # exact amounts
{"desc": "Ana pays Ben back", "amount": 50, "paid_by": "Ana", "transfer_to": "Ben"} # direct repayment
]
}
All math is done in integer cents. Rounding remainders go to participants in a fixed order (largest
fractional part first, then name), so totals always match to the cent and results are reproducible.
Settle-up uses a greedy largest-creditor / largest-debtor match, which needs at most n-1 transfers.
Usage:
python3 split_expenses.py trip.json
python3 split_expenses.py trip.json --json
python3 split_expenses.py --self-test
Exit code: 0 OK, 2 bad input.
"""
import argparse
import json
import sys
from decimal import Decimal, ROUND_HALF_UP
class SplitError(ValueError):
pass
def to_cents(amount, rate=Decimal(1)):
try:
d = Decimal(str(amount)) * Decimal(str(rate)) * 100
except Exception:
raise SplitError(f"bad amount {amount!r}")
return int(d.quantize(Decimal(1), rounding=ROUND_HALF_UP))
def fmt(cents, cur):
sign = "-" if cents < 0 else ""
return f"{sign}{abs(cents) // 100:,}.{abs(cents) % 100:02d} {cur}"
def allocate(total, weights):
"""Split integer total by weights {name: weight}; largest-remainder method, ties by name."""
wsum = sum(weights.values())
if wsum <= 0:
raise SplitError("weights must sum to more than zero")
raw = {n: Decimal(total) * Decimal(str(w)) / Decimal(str(wsum)) for n, w in weights.items()}
base = {n: int(v) for n, v in raw.items()}
left = total - sum(base.values())
order = sorted(weights, key=lambda n: (-(raw[n] - base[n]), n))
for n in order[:left]:
base[n] += 1
return base
def compute(data):
cur = data.get("currency") or "EUR"
rates = {k.upper(): v for k, v in (data.get("rates") or {}).items()}
people = data.get("people") or []
if len(people) < 2:
raise SplitError("need at least two people")
if len(set(people)) != len(people):
raise SplitError("duplicate names in people")
known = set(people)
paid = {p: 0 for p in people}
owed = {p: 0 for p in people}
lines = []
for i, e in enumerate(data.get("expenses") or [], start=1):
desc = e.get("desc") or f"expense {i}"
payer = e.get("paid_by")
if payer not in known:
raise SplitError(f"{desc}: paid_by {payer!r} is not in people")
ecur = (e.get("currency") or cur).upper()
if ecur == cur.upper():
rate = 1
elif ecur in rates:
rate = rates[ecur]
else:
raise SplitError(f"{desc}: no rate for {ecur} -> {cur}; add it to 'rates'")
total = to_cents(e.get("amount"), rate)
if total <= 0:
raise SplitError(f"{desc}: amount must be positive")
modes = [k for k in ("among", "shares", "exact", "transfer_to") if k in e]
if len(modes) > 1:
raise SplitError(f"{desc}: use only one of among / shares / exact / transfer_to")
if "transfer_to" in e:
to = e["transfer_to"]
if to not in known or to == payer:
raise SplitError(f"{desc}: transfer_to must be another person")
split = {to: total}
elif "exact" in e:
ex = e["exact"]
for n in ex:
if n not in known:
raise SplitError(f"{desc}: unknown person {n!r}")
split = {n: to_cents(v, rate) for n, v in ex.items()}
diff = total - sum(split.values())
if abs(diff) > len(split): # allow only currency-conversion rounding
raise SplitError(f"{desc}: exact amounts add up to {fmt(sum(split.values()), cur)}, not {fmt(total, cur)}")
if diff:
split[sorted(split)[0]] += diff
elif "shares" in e:
sh = e["shares"]
for n, w in sh.items():
if n not in known:
raise SplitError(f"{desc}: unknown person {n!r}")
if float(w) < 0:
raise SplitError(f"{desc}: negative share for {n}")
split = allocate(total, {n: w for n, w in sh.items() if float(w) > 0})
else:
among = e.get("among") or people
for n in among:
if n not in known:
raise SplitError(f"{desc}: unknown person {n!r}")
split = allocate(total, {n: 1 for n in among})
paid[payer] += total
for n, c in split.items():
owed[n] += c
lines.append({"desc": desc, "paid_by": payer, "amount": total, "original": f"{e.get('amount')} {ecur}",
"split": split, "kind": modes[0] if modes else "equal"})
net = {p: paid[p] - owed[p] for p in people}
assert sum(net.values()) == 0, "internal error: balances do not sum to zero"
return {"currency": cur, "people": people, "expenses": lines, "paid": paid, "owed": owed,
"net": net, "transfers": settle(net),
"total_spent": sum(l["amount"] for l in lines if l["kind"] != "transfer_to")}
def settle(net):
cred = sorted([[c, n] for n, c in net.items() if c > 0], key=lambda x: (-x[0], x[1]))
debt = sorted([[-c, n] for n, c in net.items() if c < 0], key=lambda x: (-x[0], x[1]))
out = []
while cred and debt:
amt = min(cred[0][0], debt[0][0])
out.append({"from": debt[0][1], "to": cred[0][1], "amount": amt})
cred[0][0] -= amt
debt[0][0] -= amt
cred = sorted([x for x in cred if x[0] > 0], key=lambda x: (-x[0], x[1]))
debt = sorted([x for x in debt if x[0] > 0], key=lambda x: (-x[0], x[1]))
return out
def to_markdown(r):
c = r["currency"]
lines = ["# Trip settle-up", "", f"Total spent: {fmt(r['total_spent'], c)} across {len([e for e in r['expenses'] if e['kind'] != 'transfer_to'])} expenses", "",
"## Expenses", "| # | Expense | Paid by | Amount | Split |", "|---|---|---|---|---|"]
for i, e in enumerate(r["expenses"], 1):
split = ", ".join(f"{n} {fmt(v, c)}" for n, v in sorted(e["split"].items()))
conv = "" if e["original"].endswith(c.upper()) else f" ({e['original']})"
lines.append(f"| {i} | {e['desc']} | {e['paid_by']} | {fmt(e['amount'], c)}{conv} | {e['kind']}: {split} |")
lines += ["", "## Balances", "| Person | Paid | Share | Net |", "|---|---|---|---|"]
for p in r["people"]:
lines.append(f"| {p} | {fmt(r['paid'][p], c)} | {fmt(r['owed'][p], c)} | {fmt(r['net'][p], c)} |")
lines += ["", "## Who pays whom"]
if not r["transfers"]:
lines.append("Everyone is even.")
lines += [f"- {t['from']} pays {t['to']} {fmt(t['amount'], c)}" for t in r["transfers"]]
return "\n".join(lines)
def self_test():
cases = []
def ok(name, cond):
cases.append((name, bool(cond)))
base = {"currency": "EUR", "people": ["Ana", "Ben", "Cem"]}
r = compute(dict(base, expenses=[{"desc": "Taxi", "amount": 10, "paid_by": "Ana"}]))
ok("10.00 split three ways sums exactly", sum(r["owed"].values()) == 1000 and sorted(r["owed"].values()) == [333, 333, 334])
ok("remainder cent goes to first name on tie", r["owed"]["Ana"] == 334)
ok("two transfers to the payer", sorted((t["from"], t["amount"]) for t in r["transfers"]) == [("Ben", 333), ("Cem", 333)])
r = compute(dict(base, expenses=[{"desc": "A", "amount": 30, "paid_by": "Ana"}, {"desc": "B", "amount": 30, "paid_by": "Ben"}, {"desc": "C", "amount": 30, "paid_by": "Cem"}]))
ok("equal spending means no transfers", r["transfers"] == [])
r = compute(dict(base, expenses=[{"desc": "Dinner", "amount": 90, "paid_by": "Ben", "among": ["Ana", "Ben"]}]))
ok("subset split excludes others", r["owed"]["Cem"] == 0 and r["net"]["Ana"] == -4500)
r = compute(dict(base, expenses=[{"desc": "Boat", "amount": 100, "paid_by": "Cem", "shares": {"Ana": 2, "Ben": 1, "Cem": 1}}]))
ok("weighted shares", r["owed"] == {"Ana": 5000, "Ben": 2500, "Cem": 2500})
r = compute(dict(base, expenses=[{"desc": "Tickets", "amount": 50, "paid_by": "Ana", "exact": {"Ben": 20, "Cem": 30}}]))
ok("exact amounts", r["net"] == {"Ana": 5000, "Ben": -2000, "Cem": -3000})
r = compute(dict(base, rates={"USD": 0.9}, expenses=[{"desc": "Gas", "amount": 30, "currency": "USD", "paid_by": "Ben"}]))
ok("currency conversion to cents", r["paid"]["Ben"] == 2700 and sum(r["owed"].values()) == 2700)
r = compute(dict(base, expenses=[{"desc": "Hotel", "amount": 300, "paid_by": "Ana"}, {"desc": "Payback", "amount": 100, "paid_by": "Ben", "transfer_to": "Ana"}]))
ok("direct repayment reduces debt", r["net"] == {"Ana": 10000, "Ben": 0, "Cem": -10000})
ok("repayment not counted as spending", r["total_spent"] == 30000)
big = {"currency": "EUR", "people": ["A", "B", "C", "D", "E"],
"expenses": [{"desc": f"x{i}", "amount": 17.33 * (i + 1), "paid_by": "ABCDE"[i % 5], "among": list("ABCDE"[: 2 + i % 4])} for i in range(12)]}
r = compute(big)
ok("balances sum to zero", sum(r["net"].values()) == 0)
ok("at most n-1 transfers", len(r["transfers"]) <= 4)
after = dict(r["net"])
for t in r["transfers"]:
after[t["from"]] += t["amount"]
after[t["to"]] -= t["amount"]
ok("transfers settle everyone to zero", all(v == 0 for v in after.values()))
ok("deterministic output", compute(big)["transfers"] == r["transfers"])
bad_inputs = [
(dict(base, expenses=[{"desc": "X", "amount": 10, "paid_by": "Zed"}]), "unknown payer rejected"),
(dict(base, expenses=[{"desc": "X", "amount": 10, "currency": "GBP", "paid_by": "Ana"}]), "missing rate rejected"),
(dict(base, expenses=[{"desc": "X", "amount": 10, "paid_by": "Ana", "exact": {"Ben": 3}}]), "exact mismatch rejected"),
(dict(base, expenses=[{"desc": "X", "amount": -5, "paid_by": "Ana"}]), "negative amount rejected"),
({"people": ["Solo"], "expenses": []}, "single person rejected"),
]
for data, label in bad_inputs:
try:
compute(data)
ok(label, False)
except SplitError:
ok(label, True)
ok("markdown has settle-up section", "## Who pays whom" in to_markdown(compute(dict(base, expenses=[{"desc": "Taxi", "amount": 10, "paid_by": "Ana"}]))))
passed = sum(c for _, c in cases)
for name, c in cases:
print(("PASS " if c else "FAIL ") + name)
print(f"{passed}/{len(cases)} passed")
return 0 if passed == len(cases) else 1
def main(argv=None):
ap = argparse.ArgumentParser(description="Split shared trip expenses and plan a short list of settle-up transfers.")
ap.add_argument("file", nargs="?", help="trip JSON file, or '-' for stdin")
ap.add_argument("--json", action="store_true")
ap.add_argument("--self-test", action="store_true")
a = ap.parse_args(argv)
if a.self_test:
return self_test()
if not a.file:
ap.error("trip JSON required (or --self-test)")
try:
data = json.load(sys.stdin if a.file == "-" else open(a.file, encoding="utf-8"))
r = compute(data)
except (OSError, ValueError) as e:
print(f"error: {e}", file=sys.stderr)
return 2
print(json.dumps(r, indent=2, ensure_ascii=False) if a.json else to_markdown(r))
return 0
if __name__ == "__main__":
sys.exit(main())Turns request counts into SLO answers: SLI, error budget for a 28d or 30d window, budget consumed and remaining, burn rates over 5m to 3d, multiwindow burn-rate alerts (page at 14.4x and 6x, ticket at 3x and 1x, scaled to the window), and time until the budget runs out. Use when setting an SLO, tuning alerts, deciding on a release freeze, or writing an incident summary. Includes a stdlib Python calculator with --self-test.
---
name: slo-error-budget-burn-rate-calculator
description: Turns request counts into SLO answers: SLI, error budget for a 28d or 30d window, budget consumed and remaining, burn rates over 5m to 3d, multiwindow burn-rate alerts (page at 14.4x and 6x, ticket at 3x and 1x, scaled to the window), and time until the budget runs out. Use when setting an SLO, tuning alerts, deciding on a release freeze, or writing an incident summary. Includes a stdlib Python calculator with --self-test.
---
# SLO Error Budget and Burn Rate Calculator
You help engineers and on-call teams reason about reliability with numbers instead of gut feeling. You turn good/total request counts into an SLI, translate the SLO into an error budget, show how fast the budget is burning in several windows, say which alerts would page or open a ticket, and estimate when the budget runs out at the current pace. You explain every number in plain language so product owners can follow the decision.
## Files in this skill
- `scripts/slo_budget.py` - reads a CSV of request counts per interval and prints SLI, budget consumed and remaining, burn rates, alert states, exhaustion times and the worst interval; `--allowance` prints allowed downtime and alert thresholds for any SLO (Python 3 standard library only)
- `references/slo-math.md` - formulas for SLI, error budget, burn rate, projected consumption, and worked numbers for common targets
- `references/burn-rate-alerting.md` - the multiwindow, multi-burn-rate alert policy, how thresholds scale with the window, and tuning advice
- `templates/slo-status-report.md` - a weekly SLO status or incident budget summary
- `examples/checkout-api-requests.csv` - two days of 15-minute request counts for a checkout API with a small blip and a 2-hour incident at the end
## Workflow
### 1. Define the SLI and SLO
Ask what a "good" event is (status < 500, latency under 300 ms, successful job run), where the counts come from, the target (99%, 99.5%, 99.9%, 99.95%), and the window (rolling 28 or 30 days). Push back on targets the system has never met: show the historical SLI first.
### 2. Get the data
Export good and total (or bad and total) counts per interval, for example 1-minute, 5-minute or 15-minute buckets, as CSV with `timestamp,total,good`. The timestamp marks the start of each interval. Use UTC.
### 3. Run the calculator
```
python3 scripts/slo_budget.py examples/checkout-api-requests.csv --slo 99.9 --window 30d
python3 scripts/slo_budget.py requests.csv --slo 99.5 --window 28d --json > slo.json
python3 scripts/slo_budget.py requests.csv --now 2026-10-19T22:00:00Z
python3 scripts/slo_budget.py --allowance --slo 99.95 --window 30d
python3 scripts/slo_budget.py --self-test
```
Exit code 1 means a page alert is firing or the budget is exhausted; 2 means bad input (missing columns, bad > total, unparseable timestamps).
### 4. Interpret
- Burn rate 1.0 spends exactly the budget over the window; 14.4 for an hour spends 2% of a 30-day budget.
- Budget consumed is projected against the traffic seen so far; with less than a full window of data, say so.
- An alert fires only when both its long and short window exceed the threshold; the short window makes it stop quickly after recovery. See `references/burn-rate-alerting.md`.
- Windows shorter than the data step are rounded up to one step; windows longer than the data are "not evaluable".
### 5. Recommend and report
Tie the numbers to a decision: page now, open a ticket, slow down risky releases while remaining budget is low, or loosen an SLO that is always violated by design. Fill `templates/slo-status-report.md`.
## Example result
For `examples/checkout-api-requests.csv` at 99.9% over 30d: SLI 99.8768%, 8.2% of the budget consumed, 1h burn 26.46 and 5m burn 28.06 so the fast page fires, 6h burn 5.83 just under 6.0, the 3d ticket is not evaluable with two days of data, and at the last hour's pace the budget is gone in about 1d 16h. Exit code 1.
## Rules
- State the SLO, window, data range and any missing intervals with every result.
- Do not mix good-event definitions in one SLI; compute separate SLOs for availability and latency.
- An SLO is a decision tool, not a blame tool: describe impact on users, not on people.
- Do not invent traffic numbers; if counts are missing, ask for them or say which result cannot be computed.
FILE:references/slo-math.md
# SLO Math
## Definitions
- **SLI** (service level indicator) = good events / total events, as a percentage.
- **SLO** (objective) = target SLI over a window, e.g. 99.9% of requests succeed over a rolling 30 days.
- **Error budget** = 1 - SLO = the share of events allowed to fail. 99.9% -> 0.1%.
- **Budget in events** = (1 - SLO) x total events in the window.
- **Error rate** over a window = bad / total in that window.
- **Burn rate** = error rate / (1 - SLO). 1.0 means the budget would be exactly used up at the end of the window.
## Projected consumption with partial data
With D seconds of data and W seconds of SLO window:
- expected requests in window = (requests in data / D) x W
- allowed bad = (1 - SLO) x expected requests
- consumed = bad in data / allowed bad
- time to exhaustion at a recent rate = (allowed bad - bad so far) / (bad per second in that recent window)
## Allowed full downtime
| SLO | per 30 days | per 28 days | per week | per day |
|---|---|---|---|---|
| 99% | 7h 12m | 6h 43m 12s | 1h 40m 48s | 14m 24s |
| 99.5% | 3h 36m | 3h 21m 36s | 50m 24s | 7m 12s |
| 99.9% | 43m 12s | 40m 19s | 10m 5s | 1m 26s |
| 99.95% | 21m 36s | 20m 10s | 5m 2s | 43s |
| 99.99% | 4m 19s | 4m 2s | 1m 0s | 9s |
Request-based SLOs do not care about clock time directly: 10 minutes of 50% errors costs the same as 5 minutes of full downtime at equal traffic, and an outage at 3 a.m. costs less budget than one at peak.
## Worked example
SLO 99.9%, 30d window, 1,000,000 requests per day.
- Budget: 0.001 x 30,000,000 = 30,000 failed requests.
- A 2% error rate for one hour at 41,667 req/h burns 833 failures = 2.8% of the budget, burn rate 20.
- At burn rate 20 the whole budget lasts 30d / 20 = 36 hours.
## Pitfalls
- Counting retries as separate good events hides user-facing failures.
- Very low traffic (night, new region) makes error rates noisy; use a minimum request count or a longer short window.
- Synthetic probes measure the probe, not users; label them separately.
- Changing the SLO mid-window: recompute the whole window with the new target and say so.
FILE:references/burn-rate-alerting.md
# Burn-Rate Alerting
## Default policy (multiwindow, multi-burn-rate)
| Severity | Budget share | Long window | Short window | Burn threshold (30d) | Burn threshold (28d) |
|---|---|---|---|---|---|
| page | 2% | 1h | 5m | 14.4 | 13.44 |
| page | 5% | 6h | 30m | 6.0 | 5.6 |
| ticket | 10% | 1d | 2h | 3.0 | 2.8 |
| ticket | 10% | 3d | 6h | 1.0 | 0.93 |
Threshold = budget share x SLO window / long window. Example: 0.02 x 720h / 1h = 14.4.
## Why two windows
- The **long window** proves the problem is significant (enough budget spent).
- The **short window** (about 1/12 of the long one) proves it is still happening, so the alert resets within minutes after recovery instead of staying red for an hour.
- Fire only when both exceed the threshold.
## Example alert rule (Prometheus style)
```
(
sum(rate(http_requests_errors_total[1h])) / sum(rate(http_requests_total[1h])) > (14.4 * 0.001)
and
sum(rate(http_requests_errors_total[5m])) / sum(rate(http_requests_total[5m])) > (14.4 * 0.001)
)
```
## Tuning
- Low-traffic services: require a minimum number of requests in the short window, or page on 6h only.
- Too many pages: check whether the SLO is stricter than users need before raising thresholds.
- Missed incidents: slow burns below 1.0 never alert by design; review the budget weekly.
- Data step coarser than the short window (e.g. 15-minute buckets): the short window becomes one bucket and recovery detection is slower.
## Budget policy ideas
- Over 50% consumed: review risky changes, prefer canaries.
- Over 75%: release freeze for non-fixes, reliability work first.
- Exhausted: postmortem required; freeze until the rolling window recovers.
FILE:templates/slo-status-report.md
# SLO Status - <service> - <period or incident>
**SLI:** <what counts as good> **SLO:** <target>% over rolling <28d/30d>
**Data:** <start> to <end> UTC, step <interval>, <n> missing intervals
**Calculator:** `slo_budget.py --slo <target> --window <window>` (exit <code>)
## Budget
| Metric | Value |
|---|---|
| SLI in period | |
| Budget consumed | |
| Budget remaining | |
| Allowed full downtime per window | |
## Burn rates
| 5m | 30m | 1h | 2h | 6h | 1d | 3d |
|---|---|---|---|---|---|---|
| | | | | | | |
## Alerts
| Policy | Status | Long / short burn |
|---|---|---|
| page 1h/5m | | |
| page 6h/30m | | |
| ticket 1d/2h | | |
| ticket 3d/6h | | |
## What happened
<worst interval, user impact, cause if known>
## Decision
- [ ] Page / incident open
- [ ] Release policy: <normal / caution / freeze> because <remaining budget>
- [ ] SLO or alert tuning: <change and reason>
## Summary for stakeholders
> <service> met <SLI>% against a <target>% target. <x>% of this window's error budget is used; at the current pace it lasts <time>. Next step: <decision>.
FILE:examples/checkout-api-requests.csv
timestamp,total,good
2026-10-18T00:00:00Z,4757,4755
2026-10-18T00:15:00Z,4525,4523
2026-10-18T00:30:00Z,4313,4311
2026-10-18T00:45:00Z,4122,4120
2026-10-18T01:00:00Z,3951,3949
2026-10-18T01:15:00Z,3803,3801
2026-10-18T01:30:00Z,3678,3676
2026-10-18T01:45:00Z,3577,3576
2026-10-18T02:00:00Z,3500,3498
2026-10-18T02:15:00Z,3448,3446
2026-10-18T02:30:00Z,3421,3419
2026-10-18T02:45:00Z,3019,3018
2026-10-18T03:00:00Z,3044,3043
2026-10-18T03:15:00Z,3093,3092
2026-10-18T03:30:00Z,3169,3168
2026-10-18T03:45:00Z,3270,3268
2026-10-18T04:00:00Z,3396,3394
2026-10-18T04:15:00Z,3547,3545
2026-10-18T04:30:00Z,3722,3720
2026-10-18T04:45:00Z,3921,3919
2026-10-18T05:00:00Z,4143,4141
2026-10-18T05:15:00Z,4388,4386
2026-10-18T05:30:00Z,4253,4251
2026-10-18T05:45:00Z,4539,4537
2026-10-18T06:00:00Z,4845,4843
2026-10-18T06:15:00Z,5168,5166
2026-10-18T06:30:00Z,5509,5507
2026-10-18T06:45:00Z,5865,5863
2026-10-18T07:00:00Z,6236,6234
2026-10-18T07:15:00Z,6619,6616
2026-10-18T07:30:00Z,7013,7009
2026-10-18T07:45:00Z,7418,7414
2026-10-18T08:00:00Z,7831,7827
2026-10-18T08:15:00Z,7850,7847
2026-10-18T08:30:00Z,8274,8271
2026-10-18T08:45:00Z,8702,8699
2026-10-18T09:00:00Z,9132,9127
2026-10-18T09:15:00Z,9561,9556
2026-10-18T09:30:00Z,9989,9984
2026-10-18T09:45:00Z,10413,10408
2026-10-18T10:00:00Z,10832,10827
2026-10-18T10:15:00Z,11245,11240
2026-10-18T10:30:00Z,11650,11645
2026-10-18T10:45:00Z,12044,12038
2026-10-18T11:00:00Z,12028,12022
2026-10-18T11:15:00Z,12398,12392
2026-10-18T11:30:00Z,12754,12748
2026-10-18T11:45:00Z,13095,13089
2026-10-18T12:00:00Z,13418,13412
2026-10-18T12:15:00Z,13724,13719
2026-10-18T12:30:00Z,14010,14003
2026-10-18T12:45:00Z,14275,14268
2026-10-18T13:00:00Z,14520,14513
2026-10-18T13:15:00Z,14742,14735
2026-10-18T13:30:00Z,14941,14934
2026-10-18T13:45:00Z,14716,14710
2026-10-18T14:00:00Z,14867,14861
2026-10-18T14:15:00Z,14993,14985
2026-10-18T14:30:00Z,15094,15086
2026-10-18T14:45:00Z,15170,15163
2026-10-18T15:00:00Z,15220,15213
2026-10-18T15:15:00Z,15244,15237
2026-10-18T15:30:00Z,15242,15236
2026-10-18T15:45:00Z,15215,15209
2026-10-18T16:00:00Z,15163,15155
2026-10-18T16:15:00Z,14686,14679
2026-10-18T16:30:00Z,14585,14578
2026-10-18T16:45:00Z,14460,14453
2026-10-18T17:00:00Z,14312,14306
2026-10-18T17:15:00Z,14141,14135
2026-10-18T17:30:00Z,13950,13866
2026-10-18T17:45:00Z,13738,13656
2026-10-18T18:00:00Z,13506,13499
2026-10-18T18:15:00Z,13257,13251
2026-10-18T18:30:00Z,12990,12984
2026-10-18T18:45:00Z,12708,12702
2026-10-18T19:00:00Z,12012,12007
2026-10-18T19:15:00Z,11702,11697
2026-10-18T19:30:00Z,11382,11376
2026-10-18T19:45:00Z,11051,11045
2026-10-18T20:00:00Z,10712,10707
2026-10-18T20:15:00Z,10367,10362
2026-10-18T20:30:00Z,10017,10013
2026-10-18T20:45:00Z,9663,9659
2026-10-18T21:00:00Z,9308,9304
2026-10-18T21:15:00Z,8952,8947
2026-10-18T21:30:00Z,8598,8594
2026-10-18T21:45:00Z,7848,7844
2026-10-18T22:00:00Z,7503,7500
2026-10-18T22:15:00Z,7164,7161
2026-10-18T22:30:00Z,6833,6830
2026-10-18T22:45:00Z,6513,6510
2026-10-18T23:00:00Z,6204,6201
2026-10-18T23:15:00Z,5907,5904
2026-10-18T23:30:00Z,5625,5622
2026-10-18T23:45:00Z,5358,5356
2026-10-19T00:00:00Z,5109,5107
2026-10-19T00:15:00Z,4877,4875
2026-10-19T00:30:00Z,4265,4263
2026-10-19T00:45:00Z,4074,4072
2026-10-19T01:00:00Z,3903,3901
2026-10-19T01:15:00Z,3755,3753
2026-10-19T01:30:00Z,3630,3628
2026-10-19T01:45:00Z,3529,3527
2026-10-19T02:00:00Z,3452,3451
2026-10-19T02:15:00Z,3400,3399
2026-10-19T02:30:00Z,3373,3371
2026-10-19T02:45:00Z,3371,3369
2026-10-19T03:00:00Z,3396,3394
2026-10-19T03:15:00Z,3045,3044
2026-10-19T03:30:00Z,3121,3120
2026-10-19T03:45:00Z,3222,3221
2026-10-19T04:00:00Z,3348,3347
2026-10-19T04:15:00Z,3499,3497
2026-10-19T04:30:00Z,3674,3672
2026-10-19T04:45:00Z,3873,3871
2026-10-19T05:00:00Z,4095,4093
2026-10-19T05:15:00Z,4340,4338
2026-10-19T05:30:00Z,4605,4603
2026-10-19T05:45:00Z,4491,4489
2026-10-19T06:00:00Z,4797,4795
2026-10-19T06:15:00Z,5120,5117
2026-10-19T06:30:00Z,5461,5458
2026-10-19T06:45:00Z,5817,5814
2026-10-19T07:00:00Z,6188,6185
2026-10-19T07:15:00Z,6571,6568
2026-10-19T07:30:00Z,6965,6962
2026-10-19T07:45:00Z,7370,7366
2026-10-19T08:00:00Z,7783,7779
2026-10-19T08:15:00Z,8202,8198
2026-10-19T08:30:00Z,8226,8222
2026-10-19T08:45:00Z,8654,8650
2026-10-19T09:00:00Z,9084,9080
2026-10-19T09:15:00Z,9513,9509
2026-10-19T09:30:00Z,9941,9936
2026-10-19T09:45:00Z,10365,10360
2026-10-19T10:00:00Z,10784,10779
2026-10-19T10:15:00Z,11197,11192
2026-10-19T10:30:00Z,11602,11597
2026-10-19T10:45:00Z,11996,11991
2026-10-19T11:00:00Z,12380,12375
2026-10-19T11:15:00Z,12350,12344
2026-10-19T11:30:00Z,12706,12700
2026-10-19T11:45:00Z,13047,13041
2026-10-19T12:00:00Z,13370,13364
2026-10-19T12:15:00Z,13676,13670
2026-10-19T12:30:00Z,13962,13956
2026-10-19T12:45:00Z,14227,14221
2026-10-19T13:00:00Z,14472,14464
2026-10-19T13:15:00Z,14694,14687
2026-10-19T13:30:00Z,14893,14886
2026-10-19T13:45:00Z,15068,15061
2026-10-19T14:00:00Z,14819,14812
2026-10-19T14:15:00Z,14945,14939
2026-10-19T14:30:00Z,15046,15040
2026-10-19T14:45:00Z,15122,15114
2026-10-19T15:00:00Z,15172,15164
2026-10-19T15:15:00Z,15196,15189
2026-10-19T15:30:00Z,15194,15187
2026-10-19T15:45:00Z,15167,15160
2026-10-19T16:00:00Z,15115,15109
2026-10-19T16:15:00Z,15038,15032
2026-10-19T16:30:00Z,14937,14929
2026-10-19T16:45:00Z,14412,14405
2026-10-19T17:00:00Z,14264,14257
2026-10-19T17:15:00Z,14093,14087
2026-10-19T17:30:00Z,13902,13896
2026-10-19T17:45:00Z,13690,13684
2026-10-19T18:00:00Z,13458,13453
2026-10-19T18:15:00Z,13209,13202
2026-10-19T18:30:00Z,12942,12936
2026-10-19T18:45:00Z,12660,12654
2026-10-19T19:00:00Z,12364,12358
2026-10-19T19:15:00Z,11654,11649
2026-10-19T19:30:00Z,11334,11329
2026-10-19T19:45:00Z,11003,10999
2026-10-19T20:00:00Z,10664,10658
2026-10-19T20:15:00Z,10319,10314
2026-10-19T20:30:00Z,9969,9964
2026-10-19T20:45:00Z,9615,9611
2026-10-19T21:00:00Z,9260,9256
2026-10-19T21:15:00Z,8904,8900
2026-10-19T21:30:00Z,8550,8547
2026-10-19T21:45:00Z,8200,8196
2026-10-19T22:00:00Z,7455,7298
2026-10-19T22:15:00Z,7116,6959
2026-10-19T22:30:00Z,6785,6629
2026-10-19T22:45:00Z,6465,6310
2026-10-19T23:00:00Z,6156,6002
2026-10-19T23:15:00Z,5859,5707
2026-10-19T23:30:00Z,5577,5426
2026-10-19T23:45:00Z,5310,5161
FILE:scripts/slo_budget.py
#!/usr/bin/env python3
"""SLO error budget and burn rate calculator (Python 3 standard library only).
Input: CSV with one row per interval, the timestamp marking the START of the interval, and request
counts. Accepted columns (case-insensitive): timestamp|time|ts, total|requests|count, and either
good|success|ok or bad|errors|failed. Timestamps: ISO 8601 (Z or offset) or Unix seconds.
Math (see references/slo-math.md):
error budget (fraction) = 1 - SLO e.g. 99.9% -> 0.001
error rate over window = sum(bad) / sum(total)
burn rate = error rate / (1 - SLO) 1.0 = spends exactly the budget over the SLO window
budget consumed so far = sum(bad) / ((1 - SLO) * expected requests in the SLO window)
alert threshold = budget share * window hours / long window hours (14.4 = 2% of 30d in 1h)
Default multiwindow, multi-burn-rate policy (Google SRE Workbook): page 2% in 1h (short 5m),
page 5% in 6h (short 30m), ticket 10% in 1d (short 2h), ticket 10% in 3d (short 6h). An alert fires
only when BOTH the long and the short window burn faster than the threshold. Windows shorter than
the data step are rounded up to one step; windows longer than the data are reported as not evaluable.
Usage:
python3 slo_budget.py requests.csv --slo 99.9 --window 30d
python3 slo_budget.py requests.csv --slo 99.5 --window 28d --json
python3 slo_budget.py --allowance --slo 99.95 --window 30d # downtime allowed, no CSV needed
python3 slo_budget.py --self-test
Exit code: 0 OK, 1 a page alert fires or the budget is exhausted, 2 bad input.
"""
import argparse
import csv
import datetime as dt
import json
import re
import statistics
import sys
POLICY = [
# (severity, long window, short window, budget share)
("page", "1h", "5m", 0.02),
("page", "6h", "30m", 0.05),
("ticket", "1d", "2h", 0.10),
("ticket", "3d", "6h", 0.10),
]
ALIASES = {
"ts": ("timestamp", "time", "ts", "date", "datetime"),
"total": ("total", "requests", "count", "events"),
"good": ("good", "success", "ok", "successes"),
"bad": ("bad", "errors", "failed", "failures", "error"),
}
class SLOError(ValueError):
pass
def parse_duration(s):
m = re.fullmatch(r"\s*(\d+(?:\.\d+)?)\s*([smhdw])\s*", str(s).lower())
if not m:
raise SLOError(f"bad duration {s!r} (use e.g. 5m, 1h, 28d)")
return float(m.group(1)) * {"s": 1, "m": 60, "h": 3600, "d": 86400, "w": 604800}[m.group(2)]
def fmt_dur(sec):
sec = int(round(sec))
if sec >= 86400:
d, h = divmod(sec // 3600, 24)
return f"{d}d {h}h" if h else f"{d}d"
if sec >= 3600:
h, m = divmod(sec // 60, 60)
return f"{h}h {m}m" if m else f"{h}h"
if sec >= 60:
m, s = divmod(sec, 60)
return f"{m}m {s}s" if s else f"{m}m"
return f"{sec}s"
def parse_ts(v):
v = v.strip()
if re.fullmatch(r"\d+(\.\d+)?", v):
return dt.datetime.fromtimestamp(float(v), dt.timezone.utc)
t = dt.datetime.fromisoformat(v.replace("Z", "+00:00").replace(" ", "T"))
if t.tzinfo is None:
t = t.replace(tzinfo=dt.timezone.utc)
return t.astimezone(dt.timezone.utc)
def load_rows(lines):
rd = csv.DictReader(lines)
if not rd.fieldnames:
raise SLOError("empty CSV")
cols = {}
lower = {f.strip().lower(): f for f in rd.fieldnames}
for key, names in ALIASES.items():
for n in names:
if n in lower:
cols[key] = lower[n]
break
if "ts" not in cols or "total" not in cols or not ({"good", "bad"} & cols.keys()):
raise SLOError(f"need timestamp, total and good or bad columns; got {rd.fieldnames}")
rows = []
for i, r in enumerate(rd, 2):
try:
ts = parse_ts(r[cols["ts"]])
total = int(float(r[cols["total"]]))
bad = int(float(r[cols["bad"]])) if "bad" in cols else total - int(float(r[cols["good"]]))
except (ValueError, TypeError) as e:
raise SLOError(f"line {i}: {e}")
if total < 0 or bad < 0 or bad > total:
raise SLOError(f"line {i}: counts must satisfy 0 <= bad <= total (total={total}, bad={bad})")
rows.append((ts, total, bad))
if len(rows) < 2:
raise SLOError("need at least two rows")
rows.sort(key=lambda r: r[0])
return rows
def analyze(rows, slo_pct, window, now=None, policy=POLICY):
if not 0 < slo_pct < 100:
raise SLOError("SLO must be between 0 and 100 (exclusive), e.g. 99.9")
budget = 1 - slo_pct / 100
win = parse_duration(window)
diffs = [(b[0] - a[0]).total_seconds() for a, b in zip(rows, rows[1:])]
step = statistics.median(diffs)
if step <= 0:
raise SLOError("duplicate timestamps")
first = rows[0][0]
end = (now or rows[-1][0] + dt.timedelta(seconds=step))
covered = (end - first).total_seconds()
expected_rows = int(round(covered / step))
missing = max(0, expected_rows - len(rows))
def window_stats(seconds):
seconds = max(seconds, step)
if seconds > covered + 1e-6:
return None
start = end - dt.timedelta(seconds=seconds)
tot = sum(t for ts, t, b in rows if start <= ts < end)
bad = sum(b for ts, t, b in rows if start <= ts < end)
rate = bad / tot if tot else 0.0
return {"seconds": seconds, "total": tot, "bad": bad, "error_rate": rate, "burn_rate": rate / budget}
total = sum(r[1] for r in rows)
bad = sum(r[2] for r in rows)
req_per_sec = total / covered
expected_window_total = req_per_sec * win
allowed_bad = budget * expected_window_total
consumed = bad / allowed_bad if allowed_bad else 0.0
alerts = []
for sev, lw, sw, share in policy:
thr = share * win / parse_duration(lw)
L, S = window_stats(parse_duration(lw)), window_stats(parse_duration(sw))
if L is None or S is None:
alerts.append({"severity": sev, "long": lw, "short": sw, "threshold": thr, "status": "not evaluable",
"long_burn": None, "short_burn": None})
continue
fires = L["burn_rate"] >= thr and S["burn_rate"] >= thr
alerts.append({"severity": sev, "long": lw, "short": sw, "threshold": thr,
"status": "FIRING" if fires else "ok",
"long_burn": L["burn_rate"], "short_burn": S["burn_rate"]})
burn = {}
for w in ("5m", "30m", "1h", "2h", "6h", "1d", "3d"):
s = window_stats(parse_duration(w))
burn[w] = None if s is None else s["burn_rate"]
remaining_bad = allowed_bad - bad
exhaust = {}
for w in ("1h", "6h", "1d"):
s = window_stats(parse_duration(w))
if s and s["bad"] > 0 and remaining_bad > 0:
exhaust[w] = remaining_bad / (s["bad"] / s["seconds"])
elif s:
exhaust[w] = 0.0 if remaining_bad <= 0 else None
worst = max(rows, key=lambda r: (r[2] / r[1] if r[1] else 0, r[0]))
return {
"slo_percent": slo_pct, "window": window, "window_seconds": win, "step_seconds": step,
"data_start": first.strftime("%Y-%m-%dT%H:%M:%SZ"), "data_end": end.strftime("%Y-%m-%dT%H:%M:%SZ"),
"rows": len(rows), "missing_intervals": missing,
"total": total, "bad": bad, "sli_percent": 100 * (1 - bad / total) if total else 100.0,
"allowed_bad_in_window": allowed_bad, "budget_consumed": consumed, "budget_remaining": 1 - consumed,
"allowed_downtime_seconds": budget * win,
"burn_rates": burn, "alerts": alerts, "exhaustion_seconds": exhaust,
"worst_interval": {"start": worst[0].strftime("%Y-%m-%dT%H:%M:%SZ"), "total": worst[1], "bad": worst[2],
"error_rate": worst[2] / worst[1] if worst[1] else 0.0},
}
def allowance_table(slo_pct, window):
win = parse_duration(window)
b = 1 - slo_pct / 100
return {"slo_percent": slo_pct, "window": window, "allowed_downtime_seconds": b * win,
"per_day_seconds": b * 86400, "per_week_seconds": b * 604800,
"thresholds": [{"severity": s, "long": l, "short": sh, "share": sha, "burn_rate": sha * win / parse_duration(l)}
for s, l, sh, sha in POLICY]}
def to_markdown(r):
pct = lambda x: f"{100 * x:.1f}%"
out = [f"# SLO error budget report", "",
f"**SLO:** {r['slo_percent']}% over {r['window']} **Data:** {r['data_start']} to {r['data_end']} "
f"({r['rows']} rows, step {fmt_dur(r['step_seconds'])}, {r['missing_intervals']} missing)",
f"**SLI over data:** {r['sli_percent']:.4f}% ({r['bad']:,} bad of {r['total']:,})",
f"**Error budget:** {r['allowed_bad_in_window']:,.0f} bad requests per {r['window']} at current traffic "
f"(= {fmt_dur(r['allowed_downtime_seconds'])} of full downtime)",
f"**Consumed so far:** {pct(r['budget_consumed'])} **Remaining:** {pct(r['budget_remaining'])}", "",
"## Burn rates (1.0 = on budget)", "", "| Window | Burn rate |", "|---|---|"]
for w, v in r["burn_rates"].items():
out.append(f"| {w} | {'n/a' if v is None else f'{v:.2f}'} |")
out += ["", "## Alert policy", "", "| Severity | Long / short window | Threshold | Long burn | Short burn | Status |",
"|---|---|---|---|---|---|"]
for a in r["alerts"]:
lb = "n/a" if a["long_burn"] is None else f"{a['long_burn']:.2f}"
sb = "n/a" if a["short_burn"] is None else f"{a['short_burn']:.2f}"
out.append(f"| {a['severity']} | {a['long']} / {a['short']} | {a['threshold']:.1f} | {lb} | {sb} | {a['status']} |")
out += ["", "## Time to exhaustion at recent burn"]
for w, s in r["exhaustion_seconds"].items():
out.append(f"- at last {w} rate: " + ("budget already exhausted" if s == 0 else ("no errors in window" if s is None else fmt_dur(s))))
wi = r["worst_interval"]
out += ["", f"Worst interval: {wi['start']} with {wi['bad']:,} bad of {wi['total']:,} ({100 * wi['error_rate']:.2f}%)"]
return "\n".join(out)
def self_test():
cases = []
def ok(name, cond):
cases.append((name, bool(cond)))
t0 = dt.datetime(2026, 1, 1, tzinfo=dt.timezone.utc)
def series(n, step_min, total, bad_fn):
return [(t0 + dt.timedelta(minutes=step_min * i), total, bad_fn(i)) for i in range(n)]
ok("duration parse", parse_duration("28d") == 28 * 86400 and parse_duration("5m") == 300)
ok("duration format", fmt_dur(43.2 * 60) == "43m 12s" and fmt_dur(86400 * 3) == "3d")
a = allowance_table(99.9, "30d")
ok("99.9% of 30d allows 43m 12s", round(a["allowed_downtime_seconds"]) == 2592)
ok("page threshold 14.4 for 30d", abs(a["thresholds"][0]["burn_rate"] - 14.4) < 1e-9)
ok("28d threshold scales to 13.44", abs(allowance_table(99.9, "28d")["thresholds"][0]["burn_rate"] - 13.44) < 1e-9)
flat = series(4 * 24 * 4, 15, 10000, lambda i: 10) # 0.1% errors at 99.9 -> burn 1.0
r = analyze(flat, 99.9, "30d")
ok("steady burn rate is 1.0", abs(r["burn_rates"]["1h"] - 1.0) < 1e-9 and abs(r["burn_rates"]["1d"] - 1.0) < 1e-9)
ok("no alert at burn 1.0", all(x["status"] != "FIRING" for x in r["alerts"][:3]))
ok("burn 1.0 over 4 of 30 days consumes 4/30", abs(r["budget_consumed"] - 4 / 30) < 1e-9)
ok("SLI computed", abs(r["sli_percent"] - 99.9) < 1e-9)
spike = series(4 * 24 * 4, 15, 10000, lambda i: 200 if i >= 4 * 24 * 4 - 4 else 5) # last hour 2% errors
r = analyze(spike, 99.9, "30d")
ok("fast burn pages (1h/5m)", r["alerts"][0]["status"] == "FIRING")
ok("short window rounded up to step", abs(r["burn_rates"]["5m"] - 20.0) < 1e-9)
ok("6h page not firing for 1h spike", r["alerts"][1]["status"] == "ok")
recovered = spike[:-2] + [(spike[-2][0], 10000, 5), (spike[-1][0], 10000, 5)]
r = analyze(recovered, 99.9, "30d")
ok("short window resets page after recovery", r["alerts"][0]["status"] == "ok")
short = series(8, 15, 1000, lambda i: 0)
r = analyze(short, 99.9, "30d")
ok("long windows not evaluable on 2h of data", r["alerts"][3]["status"] == "not evaluable" and r["burn_rates"]["3d"] is None)
gap = flat[:10] + flat[12:]
ok("missing intervals counted", analyze(gap, 99.9, "30d")["missing_intervals"] == 2)
ok("exhaustion time at burn 1", abs(analyze(flat, 99.9, "30d")["exhaustion_seconds"]["1d"] - 26 * 86400) < 1)
rows = load_rows(["time,requests,errors", "2026-01-01T00:00:00Z,100,1", "2026-01-01T00:05:00Z,100,0"])
ok("aliases requests/errors", rows[0][1] == 100 and rows[0][2] == 1)
rows = load_rows(["ts,total,good", "1767225600,50,49", "1767225900,50,50"])
ok("unix timestamps and good column", rows[0][2] == 1 and rows[0][0] == t0)
for data, label in [(["ts,total", "x,1"], "missing good/bad rejected"),
(["ts,total,bad", "2026-01-01T00:00:00Z,10,11", "2026-01-01T00:05:00Z,10,0"], "bad > total rejected"),
(["ts,total,bad", "nope,10,1", "2026-01-01T00:05:00Z,10,0"], "bad timestamp rejected")]:
try:
load_rows(data)
ok(label, False)
except SLOError:
ok(label, True)
try:
analyze(flat, 100, "30d")
ok("SLO 100 rejected", False)
except SLOError:
ok("SLO 100 rejected", True)
ok("markdown has alert table", "## Alert policy" in to_markdown(analyze(flat, 99.9, "30d")))
passed = sum(c for _, c in cases)
for name, c in cases:
print(("PASS " if c else "FAIL ") + name)
print(f"{passed}/{len(cases)} passed")
return 0 if passed == len(cases) else 1
def main(argv=None):
ap = argparse.ArgumentParser(description="Compute SLO error budget consumption and multiwindow burn-rate alerts.")
ap.add_argument("file", nargs="?", help="CSV of request counts per interval ('-' for stdin)")
ap.add_argument("--slo", type=float, default=99.9, help="target in percent (default 99.9)")
ap.add_argument("--window", default="30d", help="SLO window, e.g. 28d or 30d (default 30d)")
ap.add_argument("--now", help="evaluate as of this ISO time (default: end of the last interval)")
ap.add_argument("--allowance", action="store_true", help="print allowed downtime and alert thresholds only")
ap.add_argument("--json", action="store_true")
ap.add_argument("--self-test", action="store_true")
a = ap.parse_args(argv)
if a.self_test:
return self_test()
try:
if a.allowance:
t = allowance_table(a.slo, a.window)
if a.json:
print(json.dumps(t, indent=2))
else:
print(f"SLO {a.slo}% over {a.window}: {fmt_dur(t['allowed_downtime_seconds'])} of full downtime "
f"({fmt_dur(t['per_day_seconds'])}/day, {fmt_dur(t['per_week_seconds'])}/week)")
for x in t["thresholds"]:
print(f" {x['severity']:6} {x['share'] * 100:.0f}% of budget in {x['long']} (short {x['short']}): burn rate >= {x['burn_rate']:.2f}")
return 0
if not a.file:
ap.error("CSV file required (or --allowance / --self-test)")
lines = sys.stdin.read().splitlines() if a.file == "-" else open(a.file, encoding="utf-8").read().splitlines()
r = analyze(load_rows(lines), a.slo, a.window, parse_ts(a.now) if a.now else None)
except (OSError, SLOError, ValueError) as e:
print(f"error: {e}", file=sys.stderr)
return 2
print(json.dumps(r, indent=2) if a.json else to_markdown(r))
paging = any(x["severity"] == "page" and x["status"] == "FIRING" for x in r["alerts"])
return 1 if paging or r["budget_remaining"] <= 0 else 0
if __name__ == "__main__":
sys.exit(main())Checks retry, backoff and timeout settings across a call chain to prevent retry storms. Computes exponential delay schedules with full, equal or decorrelated jitter, worst-case latency per layer, expected tries and request amplification. Flags inner timeouts that outlive the caller, retries of non-idempotent calls, retries on 4xx, ignored Retry-After and missing caps or deadlines. Includes a stdlib Python calculator with --self-test.
---
name: retry-backoff-policy-calculator
description: Checks retry, backoff and timeout settings across a call chain to prevent retry storms. Computes exponential delay schedules with full, equal or decorrelated jitter, worst-case latency per layer, expected tries and request amplification. Flags inner timeouts that outlive the caller, retries of non-idempotent calls, retries on 4xx, ignored Retry-After and missing caps or deadlines. Includes a stdlib Python calculator with --self-test.
---
# Retry and Backoff Policy Calculator
You help engineers set retry, backoff and timeout policies that make services more reliable instead of turning a small outage into a big one. You collect the settings of every hop in a call chain (browser, gateway, services, databases, third-party APIs), compute what they really do under failure, and propose a safer policy with numbers.
## Files in this skill
- `scripts/retry_policy.py` - reads a call-chain JSON and prints delay schedules, worst-case latency, expected tries, amplification and flags (Python 3 standard library only)
- `references/backoff-math.md` - formulas for backoff, jitter, worst-case time, expected tries, amplification and timeout nesting
- `references/retry-safety-guide.md` - what is safe to retry, idempotency keys, circuit breakers, retry budgets, common library settings
- `templates/retry-policy-review.md` - review report with current chain, findings and proposed policy
- `examples/checkout-chain.json` - a three-layer checkout chain with typical mistakes
## Workflow
### 1. Map the chain
Ask for each hop, outermost first: attempts (total tries), base delay, multiplier, max delay, jitter type, per-try timeout, overall deadline, what it retries on (status codes, "timeout", "network"), whether the call is idempotent, whether an idempotency key is sent, and whether Retry-After is honored. Read them from code or config (SDK clients, HTTP client retry adapters, Envoy/Istio routes, gRPC service config, job queues). Unknown values: use the library default and mark them as assumed.
### 2. Run the calculator
```
python3 scripts/retry_policy.py examples/checkout-chain.json
python3 scripts/retry_policy.py chain.json --failure-rate 0.5 --max-amplification 8
python3 scripts/retry_policy.py chain.json --json > retry-report.json
python3 scripts/retry_policy.py --self-test
```
Exit code 1 means high-severity findings, 2 means bad input.
### 3. Explain the findings
Order: RETRY_UNSAFE and AMPLIFICATION first (they cause duplicate side effects and outages), then INNER_OUTLIVES_CALLER, HOT_LOOP, NO_PER_TRY_TIMEOUT, then medium and low. Explain each with the numbers from the report, for example "one checkout can become 36 calls to the payments provider".
### 4. Propose a fixed policy
Use `references/backoff-math.md` and `references/retry-safety-guide.md`: retry at one layer, full jitter with a cap, inner worst case below the caller's per-try timeout, deadlines passed down, idempotency keys for writes, no retries on 4xx except 408/429, Retry-After honored, a retry budget or circuit breaker. Re-run the script on the proposed JSON and show before and after.
### 5. Write it up
Fill `templates/retry-policy-review.md` and give config snippets for the user's library.
## Example result
For `examples/checkout-chain.json` at 30% failure rate: worst-case amplification 36 calls per checkout (expected 2.74); the gateway layer's 9.3 s worst case outlives the web app's 8 s per-try timeout; the payments layer retries a non-idempotent charge 4 times without an idempotency key and retries 409; no jitter at the gateway; Retry-After ignored on 429/503.
## Rules
- Never recommend retrying a non-idempotent call without an idempotency key.
- Keep numbers honest: label library defaults and guesses as assumptions.
- Prefer fewer retries with good timeouts over many retries.
- Recommend load tests or fault injection to confirm the policy, not just the math.
FILE:references/backoff-math.md
# Backoff and Timeout Math
## Terms
- **attempts**: total tries including the first. attempts 3 = 1 try + 2 retries.
- **base**: first retry delay. **multiplier**: growth factor (usually 2). **cap**: max single delay.
- **per-try timeout**: how long one try may take before it is abandoned.
- **deadline**: total time budget for the call including all tries and delays.
## Delay before retry k (k = 1 .. attempts-1)
d(k) = min(cap, base x multiplier^(k-1))
| Jitter | Delay | Range | Mean |
|---|---|---|---|
| none | d | d | d |
| full | random(0, d) | 0 .. d | d/2 |
| equal | d/2 + random(0, d/2) | d/2 .. d | 0.75 d |
| decorrelated | min(cap, random(base, 3 x previous)) | base .. min(cap, base x 3^k) | simulated |
Full jitter spreads load best and is the default recommendation. Equal jitter keeps a minimum wait. Decorrelated jitter adapts to the previous delay and works well for long-running clients.
## Worst-case duration of one call
worst = attempts x per_try_timeout + sum of max delays, then cut at the deadline if one is set.
Example: 3 tries x 1 s + 100 ms + 200 ms = 3.3 s.
## Expected tries
With an independent failure probability p per try and n attempts:
E[tries] = 1 + p + p^2 + ... + p^(n-1)
p = 0.3, n = 3 -> 1.39. During a full outage p = 1 and E = n.
## Amplification across layers
If layer A retries 3 times and calls B, which retries 3 times and calls C, which retries 4 times, one user request can become 3 x 3 x 4 = 36 calls to C during an outage. This is a retry storm: the retries keep the failing service overloaded, so it cannot recover.
Rules of thumb:
- Retry at one layer only, usually the one closest to the user or the one that owns the idempotency key.
- Inner layers: attempts 1 or 2 with a short deadline.
- Keep worst-case amplification at or below about 10.
- Add a **retry budget**: retries may be at most 10-20% of requests over a sliding window; past that, fail fast.
## Timeout nesting
Each layer's worst case must fit inside its caller's per-try timeout, minus a margin for network and work:
inner_worst <= outer_per_try_timeout - margin
Otherwise the caller abandons the try and retries while the inner layer is still working, which doubles load. Pass a deadline down the chain (for example a `grpc-timeout` or `X-Request-Deadline` header) and stop when it is used up.
## Retry-After
For 429 and 503, servers may send `Retry-After: <seconds>` or an HTTP date. Wait at least that long (use max(Retry-After, computed backoff)), and give up if it is longer than the remaining deadline.
FILE:references/retry-safety-guide.md
# Retry Safety Guide
## What to retry
| Outcome | Retry? | Notes |
|---|---|---|
| Connection refused / reset before the request was sent | Yes | Safe for any method |
| Timeout after the request was sent | Only if idempotent or with an idempotency key | The server may have done the work |
| 408 Request Timeout | Yes, with backoff | |
| 429 Too Many Requests | Yes, after Retry-After | Never faster than the server asks |
| 500 Internal Server Error | Sometimes | Only if the error is known to be transient |
| 502 / 503 / 504 | Yes, with backoff and jitter | Honor Retry-After on 503 |
| 400 / 401 / 403 / 404 / 405 / 410 / 422 | No | Same request fails the same way; fix the request, refresh auth once at most |
| 409 Conflict | Usually no | Exception: an idempotency key that is still in progress, retry after a short wait |
## Idempotency
- GET, HEAD, OPTIONS, PUT and DELETE are idempotent by HTTP semantics, if the server implements them correctly.
- POST and PATCH are not. To retry them, send an **idempotency key** (a UUID per logical operation, e.g. `Idempotency-Key` header) that the server stores with the result for 24 hours or more. Generate the key once, before the first try, and reuse it for every retry.
- Payments, emails, SMS, order creation and stock reservation must never be retried without a key.
## Client checklist
1. Per-try timeout set and shorter than the caller's.
2. Overall deadline set and passed down the chain.
3. Exponential backoff with a cap and full jitter.
4. attempts <= 3 at the edge, 1-2 inside.
5. Retry only on the table's "Yes" rows.
6. Idempotency key on every retried non-idempotent call.
7. Retry budget or circuit breaker so retries stop during an outage.
8. Metrics: tries per request, retry rate, give-up count, by dependency.
9. Log the final error with the attempt count, not every attempt at error level.
## Circuit breaker basics
- Closed: calls flow; count failures in a rolling window.
- Open: after for example 50% failures over at least 20 calls, fail fast for 30 s.
- Half-open: let a few trial calls through; close if they succeed.
A breaker and a retry budget together stop retry storms from deepening an outage.
## Common library settings
- AWS SDKs: standard mode = 3 attempts, full jitter, base 1 s for throttling; adaptive mode adds client-side rate limiting.
- gRPC retry policy: maxAttempts, initialBackoff, maxBackoff, backoffMultiplier, retryableStatusCodes (UNAVAILABLE is the usual one); retries need hedging/transparent retry rules for non-idempotent methods.
- Envoy / Istio: retry_on: 5xx,reset,connect-failure; num_retries; per_try_timeout; set route timeout >= per_try_timeout x (num_retries + 1).
- Python `urllib3.Retry(total=3, backoff_factor=0.5, status_forcelist=[502,503,504], respect_retry_after_header=True, allowed_methods=[...])`.
Check the library docs for the version you use; defaults change.
FILE:templates/retry-policy-review.md
# Retry Policy Review - <service / call chain> - <date>
**Reviewer:** <name> **Input:** <chain JSON / config files> **Failure rate assumed:** <30%>
## Chain
| # | Caller -> callee | Attempts | Backoff (base / x / cap) | Jitter | Per-try timeout | Deadline | Worst case |
|---|---|---|---|---|---|---|---|
| 1 | | | | | | | |
**Worst-case amplification:** <36> calls per request (target <= 10)
**Expected amplification at <p>:** <2.7>
## Findings
| Severity | Layer | Code | Problem | Fix |
|---|---|---|---|---|
| high | | | | |
## Proposed policy
| # | Caller -> callee | Attempts | Backoff | Jitter | Per-try timeout | Deadline | Retry on |
|---|---|---|---|---|---|---|---|
| 1 | | | | full | | | |
**New worst-case amplification:** <n> **New worst-case latency:** <s>
## Follow-ups
- [ ] Idempotency keys for: <calls>
- [ ] Retry budget / circuit breaker for: <dependency>
- [ ] Dashboards: tries per request, retry rate, give-ups
FILE:examples/checkout-chain.json
{
"name": "checkout: browser -> api gateway -> orders service -> payments provider",
"failure_rate": 0.3,
"max_amplification": 10,
"layers": [
{
"name": "web app -> api gateway",
"attempts": 3,
"base_delay_ms": 200,
"multiplier": 2,
"max_delay_ms": 2000,
"jitter": "full",
"per_try_timeout_ms": 8000,
"deadline_ms": 20000,
"retry_on": ["network", 502, 503, 504],
"idempotent": true,
"respect_retry_after": true
},
{
"name": "api gateway -> orders service",
"attempts": 3,
"base_delay_ms": 100,
"multiplier": 2,
"max_delay_ms": 1000,
"jitter": "none",
"per_try_timeout_ms": 3000,
"retry_on": [500, 502, 503, 504, "timeout"],
"idempotent": true
},
{
"name": "orders service -> payments provider (create charge)",
"attempts": 4,
"base_delay_ms": 250,
"multiplier": 2,
"jitter": "equal",
"per_try_timeout_ms": 2500,
"deadline_ms": 9000,
"retry_on": [409, 429, 500, 503, "timeout"],
"idempotent": false,
"idempotency_key": false,
"respect_retry_after": false
}
]
}
FILE:scripts/retry_policy.py
#!/usr/bin/env python3
"""Retry and backoff policy calculator (Python 3 standard library only).
Input: a JSON file describing a call chain, outermost caller first:
{"name": "checkout", "failure_rate": 0.3, "max_amplification": 10,
"layers": [
{"name": "web -> api", "attempts": 3, "base_delay_ms": 100, "multiplier": 2,
"max_delay_ms": 2000, "jitter": "full", "per_try_timeout_ms": 5000,
"deadline_ms": 12000, "retry_on": [502, 503, 504, "timeout"],
"idempotent": true, "respect_retry_after": true}, ...]}
attempts = total tries including the first one (attempts 1 = no retries).
jitter: none | full | equal | decorrelated. Delay before retry k (k = 1..attempts-1):
none d = min(cap, base * multiplier^(k-1))
full U(0, d) equal d/2 + U(0, d/2)
decorrelated min(cap, U(base, 3 * previous)) (AWS style; expected value simulated, seed 7)
Reports per layer: delay schedule (min / expected / max), worst-case time for one call
(attempts x per-try timeout + max delays, cut at deadline), expected tries at the given
failure rate, and for the chain: worst-case and expected request amplification at the
innermost dependency. Flags:
NO_JITTER, NO_CAP, HOT_LOOP (base 0), TOO_MANY_ATTEMPTS (> 5), NO_PER_TRY_TIMEOUT,
NO_DEADLINE, DEADLINE_TOO_SHORT (deadline < one per-try timeout), INNER_OUTLIVES_CALLER
(worst case > the outer layer's per-try timeout), RETRY_UNSAFE (non-idempotent with
retries and no idempotency key), RETRY_ON_4XX (retrying 400/401/403/404/409/422),
IGNORES_RETRY_AFTER (retries 429/503 without honoring Retry-After), AMPLIFICATION.
Usage:
python3 retry_policy.py chain.json
python3 retry_policy.py chain.json --failure-rate 0.5 --json
python3 retry_policy.py --self-test
Exit code: 0 OK, 1 high-severity flags found, 2 bad input.
"""
import argparse
import json
import random
import sys
JITTERS = ("none", "full", "equal", "decorrelated")
BAD_4XX = {400, 401, 403, 404, 405, 409, 410, 422}
SEVERITY = {"INNER_OUTLIVES_CALLER": "high", "RETRY_UNSAFE": "high", "AMPLIFICATION": "high",
"HOT_LOOP": "high", "NO_PER_TRY_TIMEOUT": "high", "RETRY_ON_4XX": "medium",
"NO_JITTER": "medium", "NO_CAP": "medium", "IGNORES_RETRY_AFTER": "medium",
"NO_DEADLINE": "medium", "DEADLINE_TOO_SHORT": "medium", "TOO_MANY_ATTEMPTS": "low"}
class PolicyError(ValueError):
pass
def num(layer, key, default=None, minimum=0):
v = layer.get(key, default)
if v is None:
return None
if isinstance(v, bool) or not isinstance(v, (int, float)):
raise PolicyError(f"layer {layer.get('name', '?')!r}: {key} must be a number")
if v < minimum:
raise PolicyError(f"layer {layer.get('name', '?')!r}: {key} must be >= {minimum}")
return v
def delays(layer):
"""Return [(min, expected, max)] in ms for each retry."""
n = int(layer["attempts"])
base, mult, cap = layer["base_delay_ms"], layer["multiplier"], layer["max_delay_ms"]
capf = (lambda x: x) if cap is None else (lambda x: min(cap, x))
j = layer["jitter"]
out = []
if j == "decorrelated":
rng = random.Random(7)
runs = 20000
sums = [0.0] * (n - 1)
for _ in range(runs):
prev = base
for k in range(n - 1):
prev = capf(rng.uniform(base, prev * 3)) if base > 0 else 0.0
sums[k] += prev
hi = base
for k in range(n - 1):
hi = capf(hi * 3)
out.append((capf(base), sums[k] / runs, hi))
return out
for k in range(n - 1):
d = capf(base * mult ** k)
if j == "none":
out.append((d, d, d))
elif j == "full":
out.append((0.0, d / 2, d))
else:
out.append((d / 2, d * 0.75, d))
return out
def normalize(raw):
layers = raw.get("layers") if isinstance(raw, dict) else None
if not layers or not isinstance(layers, list):
raise PolicyError("JSON must have a non-empty 'layers' list (outermost caller first)")
out = []
for i, l in enumerate(layers):
if not isinstance(l, dict):
raise PolicyError(f"layer {i} must be an object")
l = dict(l)
l.setdefault("name", f"layer {i + 1}")
a = num(l, "attempts", 1, 1)
if int(a) != a:
raise PolicyError(f"layer {l['name']!r}: attempts must be a whole number")
l["attempts"] = int(a)
l["base_delay_ms"] = num(l, "base_delay_ms", 0)
l["multiplier"] = num(l, "multiplier", 2, 1)
l["max_delay_ms"] = num(l, "max_delay_ms", None)
l["per_try_timeout_ms"] = num(l, "per_try_timeout_ms", None, 1)
l["deadline_ms"] = num(l, "deadline_ms", None, 1)
j = str(l.get("jitter", "none")).lower()
if j not in JITTERS:
raise PolicyError(f"layer {l['name']!r}: jitter must be one of {', '.join(JITTERS)}")
l["jitter"] = j
l["retry_on"] = list(l.get("retry_on", []))
l["idempotent"] = bool(l.get("idempotent", True))
l["idempotency_key"] = bool(l.get("idempotency_key", False))
l["respect_retry_after"] = bool(l.get("respect_retry_after", False))
out.append(l)
return out
def expected_tries(n, p):
if p >= 1:
return float(n)
return sum(p ** k for k in range(n))
def analyze(raw, failure_rate=None, max_amp=None):
layers = normalize(raw)
p = failure_rate if failure_rate is not None else raw.get("failure_rate", 0.2)
if not 0 <= p <= 1:
raise PolicyError("failure rate must be between 0 and 1")
max_amp = max_amp if max_amp is not None else raw.get("max_amplification", 10)
res = {"name": raw.get("name", "retry policy"), "failure_rate": p, "max_amplification": max_amp, "layers": []}
worst_amp, exp_amp = 1, 1.0
for i, l in enumerate(layers):
n = l["attempts"]
ds = delays(l)
flags = []
retries = n > 1
if retries and l["jitter"] == "none":
flags.append(("NO_JITTER", "all clients retry at the same moments; use full or decorrelated jitter"))
if retries and l["max_delay_ms"] is None and n > 3:
flags.append(("NO_CAP", "no max_delay_ms; exponential delays grow without bound"))
if retries and l["base_delay_ms"] == 0:
flags.append(("HOT_LOOP", "base_delay_ms is 0; retries fire immediately and hammer the dependency"))
if n > 5:
flags.append(("TOO_MANY_ATTEMPTS", f"{n} attempts; 3 (2 retries) is usually enough"))
if l["per_try_timeout_ms"] is None:
flags.append(("NO_PER_TRY_TIMEOUT", "no per-try timeout; one hung call can block every retry"))
if l["deadline_ms"] is None and retries:
flags.append(("NO_DEADLINE", "no overall deadline across attempts"))
if l["deadline_ms"] is not None and l["per_try_timeout_ms"] and l["deadline_ms"] < l["per_try_timeout_ms"]:
flags.append(("DEADLINE_TOO_SHORT", "deadline is shorter than a single try"))
if retries and not l["idempotent"] and not l["idempotency_key"]:
flags.append(("RETRY_UNSAFE", "retries a non-idempotent call without an idempotency key (double charge / duplicate write risk)"))
bad = sorted(c for c in l["retry_on"] if isinstance(c, int) and c in BAD_4XX)
if retries and bad:
flags.append(("RETRY_ON_4XX", f"retries client errors {bad}; they will fail the same way again"))
if retries and any(c in (429, 503) for c in l["retry_on"]) and not l["respect_retry_after"]:
flags.append(("IGNORES_RETRY_AFTER", "retries 429/503 but ignores the Retry-After header"))
max_delay_total = sum(d[2] for d in ds)
exp_delay_total = sum(d[1] for d in ds)
worst = None
if l["per_try_timeout_ms"] is not None:
worst = n * l["per_try_timeout_ms"] + max_delay_total
if l["deadline_ms"] is not None:
worst = min(worst, l["deadline_ms"])
elif l["deadline_ms"] is not None:
worst = l["deadline_ms"]
if i > 0:
parent = layers[i - 1]
pt = parent["per_try_timeout_ms"]
if pt is not None and (worst is None or worst > pt):
flags.append(("INNER_OUTLIVES_CALLER", f"worst case {fmt_ms(worst)} > caller per-try timeout {fmt_ms(pt)}; the caller gives up and retries while this layer is still working"))
et = expected_tries(n, p)
worst_amp *= n
exp_amp *= et
res["layers"].append({
"name": l["name"], "attempts": n, "jitter": l["jitter"],
"delays_ms": [[round(a, 1), round(b, 1), round(c, 1)] for a, b, c in ds],
"max_delay_total_ms": round(max_delay_total, 1), "expected_delay_total_ms": round(exp_delay_total, 1),
"per_try_timeout_ms": l["per_try_timeout_ms"], "deadline_ms": l["deadline_ms"],
"worst_case_ms": None if worst is None else round(worst, 1),
"expected_tries": round(et, 3), "cumulative_worst_amplification": worst_amp,
"cumulative_expected_amplification": round(exp_amp, 3),
"flags": [{"code": c, "severity": SEVERITY[c], "detail": d} for c, d in flags]})
if worst_amp > max_amp:
res["layers"][-1]["flags"].append({"code": "AMPLIFICATION", "severity": "high",
"detail": f"one top-level request can become {worst_amp} calls to the innermost dependency during an outage (limit {max_amp}); retry at one layer only or add a retry budget"})
res["worst_amplification"] = worst_amp
res["expected_amplification"] = round(exp_amp, 3)
res["high_flags"] = sum(1 for l in res["layers"] for f in l["flags"] if f["severity"] == "high")
return res
def fmt_ms(v):
if v is None:
return "unbounded"
if v >= 1000:
s = f"{v / 1000:.2f}".rstrip("0").rstrip(".")
return s + " s"
return f"{v:g} ms"
def to_markdown(r):
out = [f"# Retry policy report: {r['name']}", "",
f"Failure rate used for expected values: {r['failure_rate']:.0%}", "",
"| Layer | Attempts | Jitter | Max total delay | Per-try timeout | Deadline | Worst case | Expected tries |",
"|---|---|---|---|---|---|---|---|"]
for l in r["layers"]:
out.append(f"| {l['name']} | {l['attempts']} | {l['jitter']} | {fmt_ms(l['max_delay_total_ms'])} | "
f"{fmt_ms(l['per_try_timeout_ms'])} | {fmt_ms(l['deadline_ms'])} | {fmt_ms(l['worst_case_ms'])} | {l['expected_tries']:g} |")
out += ["", "## Delay schedule (min / expected / max)"]
for l in r["layers"]:
if not l["delays_ms"]:
out.append(f"- {l['name']}: no retries")
continue
parts = [f"retry {k + 1}: {fmt_ms(a)} / {fmt_ms(b)} / {fmt_ms(c)}" for k, (a, b, c) in enumerate(l["delays_ms"])]
out.append(f"- {l['name']}: " + "; ".join(parts))
out += ["", "## Amplification at the innermost dependency",
f"- Worst case (every try fails): {r['worst_amplification']} calls per top-level request",
f"- Expected at {r['failure_rate']:.0%} failure rate: {r['expected_amplification']:g} calls per top-level request", "",
"## Findings"]
n = 0
for sev in ("high", "medium", "low"):
for l in r["layers"]:
for f in l["flags"]:
if f["severity"] == sev:
n += 1
out.append(f"- **{sev.upper()} {f['code']}** ({l['name']}): {f['detail']}")
if not n:
out.append("- No findings.")
return "\n".join(out)
def self_test():
cases = []
def ok(name, cond):
cases.append((name, bool(cond)))
def one(**kw):
base = {"name": "x", "attempts": 3, "base_delay_ms": 100, "multiplier": 2, "max_delay_ms": 1000,
"jitter": "full", "per_try_timeout_ms": 1000, "deadline_ms": 5000, "retry_on": [503],
"respect_retry_after": True}
base.update(kw)
return base
def codes(r, i=0):
return {f["code"] for f in r["layers"][i]["flags"]}
d = delays(normalize({"layers": [one(jitter="none", attempts=5)]})[0])
ok("exponential no jitter 100,200,400,800", [x[2] for x in d] == [100, 200, 400, 800])
d = delays(normalize({"layers": [one(jitter="none", attempts=6)]})[0])
ok("cap applied at 1000", d[-1][2] == 1000 and d[-2][2] == 800)
d = delays(normalize({"layers": [one(jitter="full")]})[0])
ok("full jitter mean half", d[0] == (0.0, 50.0, 100))
d = delays(normalize({"layers": [one(jitter="equal")]})[0])
ok("equal jitter 50..100 mean 75", d[0] == (50.0, 75.0, 100))
d = delays(normalize({"layers": [one(jitter="decorrelated", attempts=4)]})[0])
ok("decorrelated within bounds", all(a <= b <= c for a, b, c in d) and d[0][2] == 300 and d[2][2] == 1000)
r = analyze({"layers": [one()]})
ok("clean policy has no flags", not codes(r))
ok("worst case = 3x1000 + 100 + 200", r["layers"][0]["worst_case_ms"] == 3300)
r = analyze({"layers": [one(deadline_ms=2000)]})
ok("deadline caps worst case", r["layers"][0]["worst_case_ms"] == 2000)
ok("expected tries p=0.5 n=3 is 1.75", expected_tries(3, 0.5) == 1.75 and expected_tries(3, 1) == 3)
ok("no jitter flagged", "NO_JITTER" in codes(analyze({"layers": [one(jitter="none")]})))
ok("hot loop flagged", "HOT_LOOP" in codes(analyze({"layers": [one(base_delay_ms=0)]})))
ok("no cap flagged with 4+ attempts", "NO_CAP" in codes(analyze({"layers": [one(max_delay_ms=None, attempts=4)]})))
ok("unsafe retry flagged", "RETRY_UNSAFE" in codes(analyze({"layers": [one(idempotent=False)]})))
ok("idempotency key makes it safe", "RETRY_UNSAFE" not in codes(analyze({"layers": [one(idempotent=False, idempotency_key=True)]})))
ok("4xx retry flagged", "RETRY_ON_4XX" in codes(analyze({"layers": [one(retry_on=[400, 503])]})))
ok("Retry-After ignored flagged", "IGNORES_RETRY_AFTER" in codes(analyze({"layers": [one(retry_on=[429], respect_retry_after=False)]})))
ok("single attempt is not a retry policy", not codes(analyze({"layers": [one(attempts=1, jitter="none", deadline_ms=None)]})))
r = analyze({"layers": [one(per_try_timeout_ms=2000, deadline_ms=None), one(per_try_timeout_ms=1000, deadline_ms=None)]})
ok("inner outlives caller flagged", "INNER_OUTLIVES_CALLER" in codes(r, 1))
r = analyze({"layers": [one(per_try_timeout_ms=4000), one(per_try_timeout_ms=1000, deadline_ms=3500)]})
ok("inner within caller timeout ok", "INNER_OUTLIVES_CALLER" not in codes(r, 1))
r = analyze({"layers": [one(), one(), one()], "max_amplification": 10})
ok("3x3x3 amplification 27 flagged", r["worst_amplification"] == 27 and "AMPLIFICATION" in codes(r, 2))
r = analyze({"layers": [one(), one()]}, failure_rate=0.5)
ok("expected amplification 1.75^2", r["expected_amplification"] == 3.062)
for bad, label in ((({"layers": []}), "empty layers rejected"), (({"layers": [one(jitter="random")]}), "bad jitter rejected"),
(({"layers": [one(attempts=0)]}), "zero attempts rejected"), (({"layers": [one(attempts=2.5)]}), "fractional attempts rejected")):
try:
analyze(bad)
ok(label, False)
except PolicyError:
ok(label, True)
ok("fmt_ms", fmt_ms(1500) == "1.5 s" and fmt_ms(250) == "250 ms" and fmt_ms(None) == "unbounded")
passed = sum(c for _, c in cases)
for name, c in cases:
print(("PASS " if c else "FAIL ") + name)
print(f"{passed}/{len(cases)} passed")
return 0 if passed == len(cases) else 1
def main(argv=None):
ap = argparse.ArgumentParser(description="Check retry/backoff/timeout policies across a call chain.")
ap.add_argument("file", nargs="?", help="chain JSON ('-' for stdin)")
ap.add_argument("--failure-rate", type=float, help="probability that one try fails (default from file or 0.2)")
ap.add_argument("--max-amplification", type=float, help="flag chains above this worst-case amplification (default 10)")
ap.add_argument("--json", action="store_true")
ap.add_argument("--self-test", action="store_true")
a = ap.parse_args(argv)
if a.self_test:
return self_test()
if not a.file:
ap.error("chain JSON required (or --self-test)")
try:
raw = json.load(sys.stdin if a.file == "-" else open(a.file, encoding="utf-8"))
r = analyze(raw, a.failure_rate, a.max_amplification)
except (OSError, PolicyError, ValueError) as e:
print(f"error: {e}", file=sys.stderr)
return 2
print(json.dumps(r, indent=2) if a.json else to_markdown(r))
return 1 if r["high_flags"] else 0
if __name__ == "__main__":
sys.exit(main())