validation-analysis

name	validation-analysis
description	Automated validation results analysis for post-commit workflow. Triggers when user requests validation analysis, code review summary, or checking background validation results. Analyzes Biome, TypeScript, and AI review reports.
version	v1.3.0
user-invocable	true
allowed-tools	Bash, Read, Grep

Post-Commit Validation Analysis

Target Token Efficiency: 75% (400 tokens → 100 tokens)

Purpose

Automated analysis of background validation results (Biome + TypeScript + AI Review) generated after commit.

Trigger Keywords

• "검증 결과"
• "검증 결과 분석"
• "validation results"
• "analyze validation"
• "check validation"
• "review results"
• "post-commit results"
• "background validation"
• "코드 검증 확인"
• "리뷰 결과"

Context

• Project: OpenManager VIBE v5.85.0
• Validation Workflow: post-commit hook (background, 5min timeout)
• Output Location: logs/validation/validation-complete-latest.md
• Components:
  • Biome: Full codebase lint + format check → logs/lint-reports/
  • TypeScript: Type check → logs/typecheck-reports/
  • AI Review: Codex → Gemini → Qwen rotation → reports/ai-review/
• Issue Tracking: reports/ai-review/.issue-tracking.json

Workflow

1. Check Summary File Existence

if [ -f "logs/validation/validation-complete-latest.md" ]; then
  echo "✅ Validation summary found"
else
  echo "⚠️ No validation summary yet. Run 'git commit' first."
  exit 0
fi

2. Read and Parse Summary

Read logs/validation/validation-complete-latest.md and extract:

• Commit hash
• Timestamp
• Report file paths
• Preview results

3. Analyze Each Component

A. Biome Results

Check for:

• ✅ Pass: No errors
• ⚠️ Warnings: < 5 acceptable
• ❌ Errors: Immediate action required

Common Issues:

• Type errors (strict mode violations)
• Unused variables
• Import organization issues

B. TypeScript Results

Check for:

• ✅ Pass: No type errors
• ❌ Fail: Type errors found

Common Issues:

• Missing type annotations
• Type mismatches
• any type usage

C. AI Review Results

Role Separation:

• 🤖 Codex/Gemini: Auto-review changed files (external AI, post-commit)
• 🧠 Claude Code: Analyze results and apply fixes to code (this Skill)

Check for:

• Review score (1-10)
• Critical findings from Codex/Gemini
• Improvement suggestions

Extract:

• 🔴 Critical issues (security, performance) → Apply fixes
• 🟡 Warnings (code quality) → Review and improve
• 🔵 Suggestions (best practices) → Consider for next commit

Claude Code Action:

1. Read Codex/Gemini review results
2. Analyze findings
3. Apply improvements to actual code
4. Re-run validation to verify fixes

4. Generate Priority Action List

Format:

📊 Validation Analysis Summary

Commit: [hash]
Date: [timestamp]

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

🔍 Biome: [✅ Pass / ⚠️ Warnings / ❌ Errors]
📝 TypeScript: [✅ Pass / ❌ Fail]
🤖 AI Review: [Score]/10

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

🎯 Priority Actions:

1. [CRITICAL] Fix TypeScript errors in X files
   → See: logs/typecheck-reports/typecheck-YYYYMMDD-HHMMSS.md

2. [HIGH] Address AI review security concerns
   → See: reports/ai-review/review-AI-YYYYMMDD-HHMMSS.md

3. [MEDIUM] Resolve Biome warnings
   → Run: npm run lint:fix

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

📁 Full Reports:
- ESLint: [path]
- TypeScript: [path]
- AI Review: [path]

💡 Next Steps:
- Fix critical issues first
- Re-run validation after fixes
- Commit improvements with --no-verify if needed

5. Smart Recommendations

If All Pass:

✅ All validations passed!
🚀 Code quality is excellent.
💡 No action required. Safe to push.

If Only Warnings:

⚠️ Minor issues detected.
✅ Safe to push if urgent.
💡 Recommended: Fix warnings in next commit.

If Errors Exist:

❌ Critical issues found!
🛑 DO NOT push to production.
💡 Required: Fix errors before deployment.

Token Optimization Strategy

Before (Manual):

User: "검증 결과 분석해줘"
Assistant: [reads summary file, reads each report, parses results, explains findings, formats output]
Tokens: ~400

After (Skill):

User: "검증 결과"
Skill: [executes workflow, generates priority list]
Tokens: ~100 (75% reduction)

Efficiency Gains:

• ❌ No need to explain validation workflow
• ❌ No need to read multiple report files
• ✅ Direct summary parsing
• ✅ Structured action list
• ✅ Smart priority sorting

Edge Cases

Case 1: Summary File Not Found

• Message: "No validation results yet. Did you commit recently?"
• Action: Wait 5 minutes after commit (background tasks running)

Case 2: Empty or Corrupted Summary

• Check: File size > 0 bytes
• Fallback: Read individual report files directly

Case 3: Reports Still Generating

• Symptom: Summary exists but report paths missing
• Message: "Validation in progress. Check again in 2-3 minutes."

Case 4: All Reports Missing

• Check: Background tasks may have timed out
• Action: Manually run validation scripts:

  bash scripts/lint/run-lint-check.sh
  bash scripts/typecheck/run-typecheck.sh
  bash scripts/code-review/auto-ai-review.sh
  

Success Criteria

• Summary file parsed successfully
• All 3 components analyzed
• Priority action list generated
• Clear next steps provided
• No manual file reading required

Integration with Other Skills

Trigger Sequence:

1. User commits code
2. post-commit hook runs (background)
3. After 5+ minutes, user asks: "검증 결과"
4. This Skill activates
5. If issues found, suggest:
   • lint-smoke - For quick re-validation
   • security-audit-workflow - If security issues detected

Related Skills

• lint-smoke - Quick re-validation after fixes
• security-audit-workflow - Deep security analysis
• playwright-triage - E2E test failure analysis
• ai-report-export - Export AI review results

Workflow Integration

Post-Commit Flow:

git commit
  ↓
.husky/post-commit → npm run hook:post-commit
  ↓
scripts/hooks/post-commit.js (cross-platform Node.js)
  ↓
Biome + TypeScript + AI Review (parallel, 5min timeout)
  ↓
create-summary.sh (aggregates results)
  ↓
logs/validation/validation-complete-latest.md
  ↓
[User requests] "검증 결과"
  ↓
This Skill analyzes and reports

Note: Windows에서는 background AI review가 스킵됩니다 (WSL/Linux only).

Changelog

• 2025-12-29: v1.3.0 - 이슈 트래킹 통합
  • .issue-tracking.json 이슈 추적 JSON 연동
  • review-issue-tracker.sh 스캔 도구 추가
  • 프로젝트 버전 v5.85.0 반영
• 2025-12-22: v1.2.0 - Cross-platform hooks support
  • Husky hooks → Node.js scripts 위임 (Windows 호환)
  • Workflow 다이어그램 업데이트
  • Windows background task 제한 사항 문서화
• 2025-12-12: v1.1.0 - Tech stack upgrade alignment
  • ESLint → Biome migration (v2.3.8)
  • AI Review rotation: Codex → Gemini → Qwen
• 2025-11-27: v1.0.0 - Initial implementation
  • Automated validation results parsing
  • Priority action list generation
  • Smart recommendations based on severity
  • 75% token efficiency improvement