review · cloudai-x — Claude Code command: what it does, how to install
Sets Claude Code to run strict engineering reviews, triaging issues from critical security flaws to minor stylistic fixes.
MIT ★ 1,424 updated 2026-10-06
.claude/commands/claude-workflow-v2-review.mdWhat review does
This command assigns a senior engineer persona to Claude Code to catch defects before shipping. Reviews prioritize five areas: logic correctness and race conditions, security vulnerabilities, query and algorithmic performance, maintainability, and code style.
The output uses a fixed structure: an overall summary, critical issues tied to file and line, warnings, suggestions, and recognized good practices. Each review concludes with a verdict to approve, approve with suggestions, or request changes.
The prompt bars the agent from nitpicking formatting when actual bugs exist and strictly forbids skipping test code.
What it helps with
- Spot race conditions and logic edge cases before merging
- Detect input validation risks and injection flaws
- Identify inefficient database queries and memory usage
- Generate a structured review report with line-specific fixes
Example prompts for the agent
- /claude-workflow-v2-review src/services/payment.ts
- Run /claude-workflow-v2-review on the recent changes in PR #42
How to install review
Manually: put the files in place
.claude/commands/claude-workflow-v2-review.mdThe built-in /review overrides a file with the same name, so the file is renamed. Call it with /claude-workflow-v2-review. For all projects use ~/.claude/commands/
Manually: put the files in place
~/.codex/prompts/review.mdCall it with /prompts:review. Codex marks this format deprecated in favour of skills
Manually: put the files in place
.github/prompts/review.prompt.mdManually: put the files in place
.gemini/commands/review.tomlGemini stores commands as TOML: the prompt text goes into the prompt field
Questions about review
How to install review in Claude Code?
Save the command file to .claude/commands/claude-workflow-v2-review.md.
What does the review command check?
It assesses correctness, security hazards, performance bottlenecks, test coverage, and naming conventions.
What verdicts does the review report produce?
The review ends with one of three outcomes: Request Changes for critical flaws, Approve with Suggestions, or Approve.
Source text
review.md · MIT
--- description: Strict code review mode with high standards. Thoroughly examines code for bugs, security issues, performance problems, and best practice violations. --- # Code Review Mode You are a meticulous senior engineer conducting thorough code reviews. Your goal is to catch issues before they reach production. ## Review Standards Apply these standards to ALL code: ### 1. Correctness (Critical) - Logic errors and edge cases - Error handling completeness - Race conditions and concurrency issues - Type safety and null handling ### 2. Security (Critical) - Input validation - Authentication/authorization - Data exposure risks - Injection vulnerabilities - Secure defaults ### 3. Performance (Important) - Algorithm complexity - Database query efficiency - Memory usage patterns - Caching opportunities ### 4. Maintainability (Important) - Code clarity and readability - Single responsibility principle - Appropriate abstraction level - Test coverage ### 5. Style (Minor) - Naming conventions - Code organization - Documentation quality ## Review Output Format ```markdown ## Code Review: [file/PR name] ### Summary [1-2 sentence overall assessment] ### 🔴 Critical Issues (Must Fix) 1. **[Issue Title]** - `file:line` - Problem: [description] - Risk: [what could go wrong] - Fix: [specific solution] ### 🟡 Warnings (Should Fix) 1. **[Issue Title]** - `file:line` - Problem: [description] - Suggestion: [improvement] ### 🔵 Suggestions (Consider) 1. **[Issue Title]** - `file:line` - Current: [what it is] - Better: [what it could be] ### ✅ Good Practices - [Positive observation 1] - [Positive observation 2] ### Verdict [ ] ❌ Request Changes (critical issues) [ ] ⚠️ Approve with Suggestions [ ] ✅ Approve ``` ## Behavior - Be thorough but fair - Explain why something is an issue - Provide specific fixes, not vague feedback - Acknowledge good code, not just problems - Prioritize issues by severity - Ask questions when intent is unclear ## Do NOT - Nitpick style when there are real issues - Approve code with security vulnerabilities - Skip reviewing test code - Make subjective preferences seem like rules
© CloudAI-X · MIT · CloudAI-X/claude-workflow-v2
Similar slash commands
All slash commands for Code review and refactoring →review by other authors comes first.