Review
Run comprehensive code review on current branch changesFrom its SKILL.md
npx -y skills add YoniChechik/claude-code-config --skill reviewAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
2 things to look at
- no licenseNo license file was found in the repository. Code published without one is not open source by default, so using it at work is a question for whoever answers licensing questions where you are.
- 0 stars0 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
3.6 KB, 783 tokens by cl100k_base, as published. Nobody here has run it
Review Mode
Run a comprehensive code review on current branch changes.
Additional review focus
"$ARGUMENTS"
If provided, the above gives optional extra constraints or focus areas for the review. By default, the skill reviews all current branch changes without needing any arguments.
Process
Subagent 1: Review
Use a subagent to carry out the review.
IMPORTANT: This agent is a reviewer only - do NOT modify any code. Report issues for the separate fix phase to handle.
Before starting the review, the subagent should:
- Read the plan file for context: Find and read
plan-*.mdin the current directory to understand the feature intent, expected changes, and architecture decisions. - The additional review focus above (if provided) gives extra constraints on what to concentrate on.
Then proceed with the full review workflow:
Step 1: Identify Changed Files
Check what files have been modified:
git status
git diff --name-only
Parse the output to get list of modified Python files.
Step 2: Deep Code Review
Review each modified file for:
Security Concerns:
- SQL injection vulnerabilities
- Unsafe data handling
- Credential exposure
- Input validation issues
Code Quality:
- Functions over 50 lines (should be broken down)
- Duplicated code patterns
- Missing type annotations
- Poor naming conventions
- Spaghetti code / complex control flow
Integration Issues:
- Breaking changes to existing APIs
- Missing error handling
- Race conditions
- Resource leaks
Edge Cases:
- Null/None handling
- Empty collections
- Boundary conditions
- Error scenarios
For each issue found, provide:
- File path and line number
- Severity (BLOCKING, HIGH, MEDIUM, LOW)
- Detailed explanation
- Suggested fix
Step 3: Generate Review Report
Create review.md with the following structure:
# Code Review Report
**Date**: [Current date]
**Branch**: [Branch name]
**Reviewer**: Code Review Agent
## Summary
[High-level summary of changes and overall assessment]
**Overall Status**: APPROVED / CHANGES REQUESTED / REJECTED
## Code Review Findings
### BLOCKING Issues
[Critical issues - these MUST be fixed]
### High Priority
[Security concerns, major quality issues]
### Medium Priority
[Code quality improvements, refactoring suggestions]
### Low Priority / Suggestions
[Nice-to-have improvements, style suggestions]
## Test Results
[Test run summary and any failures]
## Files Reviewed
[List of all files checked with brief notes]
Write the report to review.md in the current directory.
Subagent 2: Codex critique pass
After the review subagent writes review.md, get a second opinion from Codex:
- Critique: Invoke the
/codexskill on the current branch diff (per its PR diff review recipe). Ask it to find things the primary review missed — bugs, security issues, design smells, untested paths, missed edge cases. - Triage: Separate valid Codex findings from noise.
- Merge: Append the valid new findings into
review.mdunder the appropriate severity sections, tagged as(Codex)so the fix phase picks them up.
The skill ends with the merged review.md. Do NOT fix anything — fixing is a separate, single-writer phase.
Important Notes
- Be thorough and skeptical - better to catch issues now than in production
- Provide specific feedback - include file:line references
- Explain the "why" - don't just say what's wrong, explain why it matters
- Be constructive - suggest fixes, not just criticism
What ships with it
Read from the repository
Just SKILL.md. No reference files, no scripts.
Gives 1 of the 12 instructions most review quality skills give in 783 tokens
Counted across 1,048 of the 1,783 authors here whose files we hold, read 2026-08-07
- Ask questions one at a timein 81 of 1048, across 64 files
- Provide a recommended answer for each questionin 73 of 1048, across 50 files
- Explore the codebase instead of asking answerable questionsin 66 of 1048, across 42 files
- Resolve dependencies between decisions one-by-onein 42 of 1048, across 17 files
- Interview the user relentlessly about the planin 38 of 1048, across 13 files
- Order findings by severityin 31 of 1048
- Resolve each branch of the decision treein 27 of 1048, across 5 files
- Run a grilling sessionin 26 of 1048, across 5 files
- Update CONTEXT.md immediately when a term is resolvedin 26 of 1048, across 11 files
- Propose precise canonical terms for vague languagein 25 of 1048, across 7 files
- Create documentation files lazilyin 24 of 1048, across 5 files
- Assign severity to every findinghere, and in 24 of 1048
Said here and by no other author read
- flag functions over 50 lines
- check for breaking API changes
- write the report to review.md
- invoke the codex skill on the diff
- triage valid codex findings
- merge valid codex findings into review.md
Grouped from the skills themselves: near-identical wordings counted once, and counted by distinct author, so one author publishing three of these counts once. Length counted with cl100k_base; the agent that loads this file may tokenize it differently.