| name | verify-style-a |
| type | action |
| description | Code cleanliness — AI-generated slop, naming, dead code, over-engineering, simplification |
| user-invocable | false |
Action skill — Code cleanliness review: AI slop detection, naming, dead code, fix protocol, output format.
Code Style and Cleanliness
Line-level code quality without changing behavior.
Slop (AI-Generated Cruft)
Defensive code — Try/catch around code that can't fail. Null checks on values that are never null. Fallbacks for impossible cases. Validation of internal values guaranteed by the type system.
Over-engineering — Abstractions for one-time operations. Interfaces with single implementations. Factory patterns for simple construction. Configuration for things that will never be configured.
Verbosity — Multiple lines that could be one. Intermediate variables that add nothing. Obvious comments that restate the code.
Messy Code
Naming — Unclear names. Inconsistent naming across similar modules. Abbreviations that aren't obvious. Names that lie.
Structure — Functions doing multiple things. Files with mixed responsibilities. Duplicated logic. Complex conditionals that could be simplified.
Dead code — Unused imports. Unreachable branches. Functions nothing calls. Commented-out code.
Fix Protocol
Find, fix, test. One issue at a time. If tests fail, revert and note why.
Not architecture review, not larp detection. Don't change behavior. "No issues found" is valid.
Output Format
## Fixes Applied
### 1. {path}:{line} - {category}
BEFORE: {old code}
AFTER: {new code}
REASON: Why this was changed
## Skipped (tests failed)
- {path}:{line} - what you tried, why it broke tests
## Summary
- Fixed: N issues
- Skipped: M issues (would break tests)
- Lines removed: K
