AI code review
SlopCop uses deterministic analysis to find concrete leads, then gives AI a bounded set of source evidence to explain what matters. Every model-authored claim must point back to supplied evidence.
Static analysis identifies candidate problems such as expensive work inside loops, risky evaluation, weak assertions, complexity hotspots, credential-like patterns, and selected unsafe data flows.
Those findings are leads, not automatic verdicts. A match can be useful without proving that the surrounding program is wrong.
Quick review explains selected analyzer findings using structured evidence and nearby source excerpts. Deep review can add a bounded selection of source, documentation, architecture samples, and available recent commit metadata.
SlopCop does not ask a model to make unsupported repository-wide claims from an unbounded code dump. Invalid citations are rejected, and unavailable interpretation does not erase the underlying analyzer results.
Reports distinguish observed evidence from interpretation and suggested next steps. Heuristic findings remain candidates to verify, and incomplete checks are not presented as passes.
The result is a review you can inspect, challenge, and hand to a developer or coding agent without losing the trail back to the code.
