wiki-researcher
Conduct Deep Codebase Research with Evidence-Based Analysis
Shallow code reviews lead to misunderstandings and bugs. This skill enforces rigorous, evidence-based investigation of codebases through structured multi-turn 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
「wiki-researcher」を使用しています。 How does the caching layer work?
期待される結果:
HIGH Confidence: CacheManager (src/cache/manager.ts:15) implements LRU eviction. Cache entries flow through RedisCacheAdapter (src/cache/redis.ts:42) which calls Redis SETEX with TTL. Default TTL is 300s (config.ts:78). Cache miss fallback to DataLoader (src/dataloader/index.ts:12) for batch fetching.
「wiki-researcher」を使用しています。 What handles payment failures?
期待される結果:
MEDIUM Confidence: PaymentService.processPayment (src/services/payments.ts:89) wraps Stripe calls in try-catch. Failures log to Sentry (line 112) and emit PAYMENT_FAILED event (line 115). Event consumers: EmailService sends notification (src/email/handlers.ts:78), webhook dispatcher triggers external callbacks (src/webhook/dispatcher.ts:45). Have not verified retry logic.
セキュリティ監査
安全Static analysis flagged 5 high-risk patterns and 1 medium-risk pattern. All findings are false positives caused by markdown syntax misinterpretation (table pipes, list dashes). The skill is a legitimate codebase research tool with no network requests, file writes, or credential access. It provides structured methodology for deep code analysis using standard read-only tools (Glob, Grep, Read).
中リスクの問題 (1)
低リスクの問題 (2)
品質スコア
作れるもの
Understand Legacy Codebases
New team members rapidly learn complex systems by tracing actual code paths rather than relying on outdated documentation.
Architectural Decision Analysis
Evaluate design patterns, coupling, and technical debt before refactoring or adding new features.
Bug Investigation
Trace data flow and function calls to identify root causes of complex bugs spanning multiple modules.
これらのプロンプトを試す
How does authentication work in this codebase?
Map the data flow from API entry point to database storage for user creation.
What design patterns are used in the payment processing module? Identify trade-offs and risks.
Trace the complete request lifecycle from ingress through middleware, services, and data layers. Include all transformations, error handling, and external integrations.
ベストプラクティス
- Provide specific topics or questions rather than generic 'analyze everything' requests
- Use when you need depth rather than quick answers - this skill prioritizes thoroughness over speed
- Combine with file navigation tools to jump to specific code locations mentioned in findings
- Request Mermaid diagrams when visualizing complex architectures or data flows
回避
- Avoid using for simple grep searches - use basic file search instead
- Do not expect runtime behavior analysis - this skill traces static code paths only
- Avoid asking about files outside the accessible workspace directory
- Do not use for time-sensitive tasks - the 5-iteration process requires multiple tool calls