slopgate

Catch AI slop before it ships.

$ npx github:ceocxx/slopgate scan .

What it catches

Stubs & not implemented code paths AI agents leave behind
TODO / FIXME / placeholder comments in shipped code
Fake data: Math.random() metrics, mockData in non-test files
Swallowed errors: empty catch {} blocks
Placeholder values: YOUR_API_KEY, example.com

Real output

$ slopgate scan examples/slop-sample.js
examples/slop-sample.js
  22:20   high    not-implemented   Stub — this code path is explicitly not implemented.
  28:5    medium  empty-catch       Empty catch block silently swallows errors.
  12:21   medium  fake-randomness   Math.random() — confirm this is not fake data.

Summary: 11 finding(s) — 1 high, 8 medium, 2 low
✗ FAIL — slop found at or above threshold (medium)

Gate only what changed

Adopting on an existing repo? --diff flags slop only on the lines a change touches, so pre-existing debt doesn't block you.

$ slopgate scan . --diff origin/main

Drop it into CI

# .github/workflows/slop.yml
- uses: ceocxx/slopgate@v0.3.0
  with:
    fail-on: medium