Review

{
  "schema_version": "2.0",
  "meta": {
    "generated_at": "2026-01-16T20:05:36.386Z",
    "slug": "chachamaru127-review",
    "source_url": "https://github.com/Chachamaru127/claude-code-harness/tree/main/skills/review",
    "source_ref": "main",
    "model": "claude",
    "analysis_version": "3.0.0",
    "source_type": "community",
    "content_hash": "44c8e51632fce706b56b8bb325e820d3009616728224363c9350972d86f17405",
    "tree_hash": "7a40a965467ac2861f6e8585585049154b7a91768f5198ed5d5ac085b67c4421"
  },
  "skill": {
    "name": "review",
    "description": "Reviews code for quality, security, performance, and accessibility issues. Use when user mentions レビュー, review, コードレビュー, セキュリティ, パフォーマンス, 品質チェック, セルフレビュー, PR, diff, 変更確認. Do NOT load for: 実装作業, 新機能開発, バグ修正, セットアップ.",
    "summary": "Reviews code for quality, security, performance, and accessibility issues. Use when user mentions レビ...",
    "icon": "🔍",
    "version": "1.0.0",
    "author": "Chachamaru127",
    "license": "MIT",
    "category": "coding",
    "tags": [
      "code review",
      "quality",
      "security",
      "performance",
      "accessibility"
    ],
    "supported_tools": [
      "claude",
      "codex",
      "claude-code"
    ],
    "risk_factors": [
      "external_commands",
      "filesystem"
    ]
  },
  "security_audit": {
    "risk_level": "safe",
    "is_blocked": false,
    "safe_to_publish": true,
    "summary": "Documentation-only skill providing review frameworks. Contains markdown files with checklists and examples. No executable code. Uses read-only tools (Read, Grep, Glob, Bash, Task) only. All static findings are false positives from markdown code blocks and YAML frontmatter.",
    "risk_factor_evidence": [
      {
        "factor": "external_commands",
        "evidence": [
          {
            "file": "review-changes/doc.md",
            "line_start": 67,
            "line_end": 82
          }
        ]
      },
      {
        "factor": "filesystem",
        "evidence": [
          {
            "file": "SKILL.md",
            "line_start": 1,
            "line_end": 10
          }
        ]
      }
    ],
    "critical_findings": [],
    "high_findings": [],
    "medium_findings": [],
    "low_findings": [],
    "dangerous_patterns": [],
    "files_scanned": 7,
    "total_lines": 1279,
    "audit_model": "claude",
    "audited_at": "2026-01-16T20:05:36.386Z"
  },
  "content": {
    "user_title": "Review code for quality, security, performance, accessibility",
    "value_statement": "Manual code review is time-consuming and inconsistent. This skill provides structured review checklists and automated analysis across multiple dimensions. Developers get consistent, comprehensive feedback on every change.",
    "seo_keywords": [
      "Claude Code code review skill",
      "automated code review",
      "code quality check",
      "security vulnerability scan",
      "performance optimization review",
      "accessibility a11y checker",
      "PR review automation",
      "self-code review",
      "Claude Codex review",
      "OWASP security checklist"
    ],
    "actual_capabilities": [
      "Analyze code changes using git diff commands",
      "Check for security vulnerabilities (SQL injection, XSS, hardcoded secrets)",
      "Review performance patterns (N+1 queries, memoization, efficient algorithms)",
      "Audit accessibility compliance (WCAG, ARIA, keyboard navigation)",
      "Evaluate code quality (naming, structure, error handling, type safety)",
      "Score reviews with A-F grading system and severity levels"
    ],
    "limitations": [
      "Automated checks only - cannot replace human judgment for complex architectural decisions",
      "Does not execute code or run tests - static analysis only",
      "Context awareness depends on provided repository information",
      "Language support varies by check type (TypeScript/JavaScript focused)"
    ],
    "use_cases": [
      {
        "target_user": "Software developers",
        "title": "Pre-commit self-review",
        "description": "Review your own changes before committing to catch issues early."
      },
      {
        "target_user": "Security engineers",
        "title": "Security-focused PR review",
        "description": "Scan pull requests for OWASP Top 10 vulnerabilities and credential leaks."
      },
      {
        "target_user": "QA engineers",
        "title": "Comprehensive code audit",
        "description": "Evaluate code changes across quality, security, performance, and accessibility."
      }
    ],
    "prompt_templates": [
      {
        "title": "Quick security scan",
        "scenario": "Security review for a PR",
        "prompt": "Use the review-security skill to check src/api/ for security issues."
      },
      {
        "title": "Performance check",
        "scenario": "Performance review for data processing code",
        "prompt": "Use the review-performance skill to analyze src/utils/data-processing.ts."
      },
      {
        "title": "Accessibility audit",
        "scenario": "UI component accessibility review",
        "prompt": "Use the review-accessibility skill to audit src/components/*.tsx."
      },
      {
        "title": "Full multi-dimensional review",
        "scenario": "Comprehensive PR review",
        "prompt": "Review the changes across all dimensions: quality, security, performance, and accessibility."
      }
    ],
    "output_examples": [
      {
        "input": "Review my recent changes in src/services/user.ts for quality and security issues",
        "output": [
          "## Review Results: src/services/user.ts",
          "**Overall Score: B**",
          "",
          "### Security Issues (1)",
          "- Medium: Line 45 - SQL injection risk in query construction",
          "  Recommendation: Use parameterized queries or ORM",
          "",
          "### Quality Issues (2)",
          "- Low: Line 23 - Function exceeds 50 lines (67 lines detected)",
          "- Low: Line 12 - Generic variable name 'data' should be more descriptive"
        ]
      }
    ],
    "best_practices": [
      "Run reviews before committing to catch issues early and reduce review cycles",
      "Use the quality gate to focus on the most relevant review dimension for your changes",
      "Combine with linters (ESLint, TypeScript) for automated pre-review checks"
    ],
    "anti_patterns": [
      "Skipping reviews for small changes - many bugs hide in minor modifications",
      "Ignoring low-severity issues - they accumulate and degrade code quality over time",
      "Using reviews as punishment - frame them as learning opportunities for the team"
    ],
    "faq": [
      {
        "question": "Which languages does this skill support?",
        "answer": "Primary focus is TypeScript/JavaScript. Security checks work on any language. Performance and quality checks are most effective with typed languages."
      },
      {
        "question": "Does this skill execute the code it reviews?",
        "answer": "No, this is a static analysis tool. It reads code and runs read-only commands like git diff. It does not execute or modify any code."
      },
      {
        "question": "How accurate are the security vulnerability detections?",
        "answer": "The skill identifies common patterns but may produce false positives. Always verify findings manually, especially for complex security issues."
      },
      {
        "question": "Can this skill integrate with my CI/CD pipeline?",
        "answer": "The skill provides review commands you can run in CI. It does not directly integrate but can be invoked as part of your pipeline scripts."
      },
      {
        "question": "Is my code and data safe when using this skill?",
        "answer": "Yes. The skill only reads files from your repository using Claude Code tools. No code or data is sent externally or stored permanently."
      },
      {
        "question": "How does this compare to automated tools like SonarQube?",
        "answer": "This skill provides AI-powered contextual review with explanations. It complements tools like SonarQube, which focus on specific metrics and rules."
      }
    ]
  },
  "file_structure": [
    {
      "name": "review-accessibility",
      "type": "dir",
      "path": "review-accessibility",
      "children": [
        {
          "name": "doc.md",
          "type": "file",
          "path": "review-accessibility/doc.md",
          "lines": 147
        }
      ]
    },
    {
      "name": "review-changes",
      "type": "dir",
      "path": "review-changes",
      "children": [
        {
          "name": "doc.md",
          "type": "file",
          "path": "review-changes/doc.md",
          "lines": 216
        }
      ]
    },
    {
      "name": "review-performance",
      "type": "dir",
      "path": "review-performance",
      "children": [
        {
          "name": "doc.md",
          "type": "file",
          "path": "review-performance/doc.md",
          "lines": 128
        }
      ]
    },
    {
      "name": "review-quality",
      "type": "dir",
      "path": "review-quality",
      "children": [
        {
          "name": "doc.md",
          "type": "file",
          "path": "review-quality/doc.md",
          "lines": 134
        }
      ]
    },
    {
      "name": "review-security",
      "type": "dir",
      "path": "review-security",
      "children": [
        {
          "name": "doc.md",
          "type": "file",
          "path": "review-security/doc.md",
          "lines": 125
        }
      ]
    },
    {
      "name": "SKILL.md",
      "type": "file",
      "path": "SKILL.md",
      "lines": 227
    }
  ]
}