Complex tasks require diverse expertise that single agents cannot provide. This skill enables coordinated multi-agent workflows through Claude Code's native agent system for comprehensive analysis.
スキルZIPをダウンロード
Claudeでアップロード
設定 → 機能 → スキル → スキルをアップロードへ移動
オンにして利用開始
テストする
「parallel-agents」を使用しています。 Run a comprehensive review of the user authentication system
期待される結果:
- Security-auditor identified missing rate limiting on login endpoint
- Backend-specialist flagged inconsistent error handling across auth routes
- Test-engineer found 40% coverage gap in password reset flow
- Synthesis: Critical rate limiting fix needed, error handling standardization recommended, additional test coverage required
「parallel-agents」を使用しています。 Analyze the React component library for accessibility issues
期待される結果:
- Frontend-specialist identified 12 components missing ARIA labels
- Explorer-agent mapped 47 interactive components across 8 directories
- Recommendations: Add ARIA labels to form inputs, implement keyboard navigation for dropdowns, add screen reader announcements for dynamic content
セキュリティ監査
安全This skill is documentation-only content teaching users how to orchestrate Claude Code's built-in agents. Static analyzer flagged 32 false positives for markdown code block delimiters (```) misidentified as Ruby shell backticks, and 3 false positives for 'weak cryptography' in YAML descriptions and table cells. No executable code exists in this skill. All findings dismissed as false positives.
品質スコア
作れるもの
Comprehensive Code Review
Run security-auditor, backend-specialist, frontend-specialist, and test-engineer agents in sequence to review a pull request from multiple perspectives, then synthesize all findings into a single actionable report.
Feature Implementation
Orchestrate database-architect for schema design, backend-specialist for API implementation, frontend-specialist for UI components, and test-engineer for coverage verification in a coordinated workflow.
Security Audit Pipeline
Execute security-auditor for configuration review, penetration-tester for vulnerability assessment, and synthesize prioritized remediation steps for development teams.
これらのプロンプトを試す
Use the security-auditor agent to review the authentication module for common vulnerabilities and report findings.
First, use the explorer-agent to map the project structure. Then, have the backend-specialist review all API endpoints. Finally, ask the test-engineer to identify test coverage gaps based on those findings.
Run the security-auditor, performance-optimizer, and test-engineer agents in parallel on the checkout module. Synthesize all findings into a prioritized action plan with critical issues first.
Resume the previous agent session and continue with additional requirements: add mobile compatibility review using the mobile-developer agent and update the synthesis report accordingly.
ベストプラクティス
- Follow logical agent order: discovery first, then analysis, implementation, and testing last
- Pass relevant context between agents to avoid redundant analysis and maintain coherence
- Always include a synthesis step to consolidate findings into actionable recommendations
回避
- Using multiple agents for simple single-domain tasks that one agent can handle efficiently
- Running agents in parallel without planning how to synthesize their potentially conflicting outputs
- Skipping the test-engineer agent when implementing code modifications