reviewing-code-multi-ailisted

# Multi-AI Code Review Orchestrates multiple AI reviewers for comprehensive code analysis. ## Reviewers | Reviewer | Strength | Focus | |----------|----------|-------| | **Claude** | Logic, architecture | Design patterns, edge cases | | **Codex** | Code quality | Bugs, optimizations | | **Gemini** | Documentation | Readability, comments | | **Droid** | Security | Vulnerabilities, best practices | ## Quick Start ```bash # Review staged changes git diff --cached > /tmp/diff.txt # Run reviewers in parallel (background) # Each saves to .context/reviews/{reviewer}.md ``` ## Workflow ### Step 1: Collect Changes ```bash # For PR review gh pr diff <number> > /tmp/changes.diff # For local changes git diff HEAD~1 > /tmp/changes.diff ``` ### Step 2: Run Reviewers Each reviewer analyzes from their perspective and saves: - `.context/reviews/claude.md` - `.context/reviews/codex.md` - `.context/reviews/gemini.md` ### Step 3: Merge Results Combine all reviews into unified report: - Critical issues (all reviewers agree) - Suggestions (reviewer-specific) - Approved aspects ## Output Format ```markdown ## Code Review Summary ### 🔴 Critical (consensus) - [file:line] Issue description ### 🟡 Suggestions - **Claude**: Architecture improvement - **Codex**: Performance optimization - **Gemini**: Documentation needed ### ✅ Approved - Error handling looks good - Test coverage adequate ``` ## Best Practices - Use 2-3 reviewers for balance - Run in background for large diffs - Prio