slop-scan

/slop-scan — Detect Technical Debt

Usage

/slop-scan              # recent changes (uncommitted + staged)
/slop-scan --commits 3  # last N commits
/slop-scan src/         # changed files in a directory
/slop-scan --full       # whole project
/slop-scan --quick      # top issues only, still recent-changes unless --full

What It Detects

1. Code smells: functions >50 lines, files >500 lines, nesting >3 levels
2. Duplication: copy-pasted blocks, repeated patterns
3. Dead code: unused functions, commented-out code >10 lines, unused imports
4. Tech debt: old TODOs, deprecated APIs, outdated deps
5. Anti-patterns: magic numbers, tight coupling, missing error handling
6. AI slop: over-abstraction, unnecessary helpers, verbose comments on obvious code

Process

1. Select files (default = diff mode):
   • Default: git diff --name-only HEAD + git diff --cached --name-only
   • --commits N: git diff --name-only HEAD~N
   • --full: all files by type (exclude node_modules, .git, dist, vendor)
   • If no git changes found, ask before running --full
2. Prioritize most-changed files (git log --format=%H -- FILE | wc -l)
3. Grep for patterns (TODO, FIXME, unused, deprecated)
4. Read top hotspot files for deeper analysis
5. Score and rank by impact

Output

## Slop Report: [project]

**Slop Level**: Low/Medium/High
**Files scanned**: X | **Issues**: Y

### Critical (fix now)
| Issue | Location | Effort |
|-------|----------|--------|

### Quick Wins (low effort, high impact)
- [list with estimated time]

### Deep Refactors (plan for sprint)
- [list with estimated time]

### Hotspots (most problematic files)
1. file.ts — X issues

Decision Rules

• High (>30 issues): pause features, refactor sprint
• Medium (10-30): weekly cleanup sessions
• Low (<10): maintenance mode