What AI slop costs if you fix it later.
Every swallowed exception or leftover helper is cheapest to remove in the pull request that added it. Enter your numbers to see how many findings pile up each month and what deferring them costs.
New AI-slop findings your team adds to the codebase each month.
to fix them at merge
to fix them after they ship
saved per year by fixing at merge
A directional estimate from your inputs, not a measurement. The gap between the two fix times is the assumption that matters most, so try a few values.
No hidden math.
Findings scale with PR volume, the AI-assisted share and findings per PR. The two fix times turn the same findings into hours, and the difference between them is the cost of deferring.
- 01
findings per month = PRs per week × (AI share ÷ 100) × findings per AI PR × 4.33 - 02
hours at merge per month = findings per month × minutes to fix at merge ÷ 60 - 03
hours later per month = findings per month × minutes to fix after shipping ÷ 60 - 04
hours saved per year = (hours later − hours at merge) × 12
Stop estimating. Measure it.
aislop scores your repo against the same deterministic rules on every change and every PR. Run the local CLI, then use a hosted plan to enforce the result across your team.
npx aislop scan