| name | anti-slop-review |
| description | Self-review ThreatForge changes for AI engineering slop without removing behavior |
Anti-slop review
Run after implementation and before any completion claim or PR handoff.
- Read
AGENTS.md and docs/quality/agentic-slop.md.
- Inspect the complete diff and every changed file in context.
- Compare the result with the issue or committed plan.
- Check architecture/code, tests, docs, and operational surfaces for documented recognition
patterns.
- Protect real functionality, validation, accessibility, compatibility, and boundary defense.
- Fix
must-fix and should-fix findings when the fix is behavior-preserving and in scope.
- Raise rather than guess when a fix changes product intent or scope.
- Re-run the smallest checks that prove each correction.
Do not manufacture findings. Report:
- findings fixed
- findings raised for owner decision
- targeted verification rerun
clean, minor, or needs-work