Code review
harness-evalFrom the repository description
npx -y skills add whchoi98/harness-eval --skill code-reviewAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
One thing to look at
- 19 stars19 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
SKILL.md
1.5 KB, as published. Nobody here has run it
Code Review Skill
Review changed code with confidence-based scoring to filter false positives.
Review Scope
By default, review unstaged changes from git diff. The user may specify different files or scope.
Review Criteria
Project Guidelines Compliance
- Import patterns and module boundaries
- Framework conventions and language style
- Function declarations and error handling
- Naming conventions from CLAUDE.md
Bug Detection
- Logic errors and null/undefined handling
- Race conditions and memory leaks
- Security vulnerabilities (OWASP Top 10)
- Performance problems
Code Quality
- Code duplication and unnecessary complexity
- Missing critical error handling
- Test coverage gaps
- Accessibility issues (for frontend code)
Confidence Scoring
Rate each issue 0-100:
- 0-24: Likely false positive or pre-existing issue. Do not report.
- 25-49: Might be real but possibly a nitpick. Do not report.
- 50-74: Real issue but minor. Report only if critical.
- 75-89: Verified real issue, important. Report with fix suggestion.
- 90-100: Confirmed critical issue. Must report.
Only report issues with confidence >= 75.
Output Format
For each issue:
[CRITICAL|IMPORTANT] <issue title> (confidence: XX)
File: path/to/file.ext:line
Issue: Clear description of the problem
Guideline: Reference to CLAUDE.md rule or security standard
Fix: Concrete code suggestion
If no high-confidence issues found, confirm code meets standards with brief summary.