wiki-qa
Get instant answers about your codebase
Stop hunting through files to understand how your code works. This skill reads your source code and provides evidence-based answers with inline citations to specific files and line numbers.
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-qa」を使用しています。 How does the data validation work?
期待される結果:
Data validation is handled through a multi-layer approach:
1. **Schema Validation** - Input is validated against JSON schemas (src/validators/schema.ts:15-42)
2. **Business Rules** - Custom validators check domain constraints (src/validators/business.ts:8-25)
3. **Sanitization** - User input is sanitized before storage (src/utils/sanitize.ts:12-18)
Key Files:
| File | Role |
|------|------|
| src/validators/schema.ts | Schema definition and validation |
| src/validators/business.ts | Domain-specific rules |
| src/utils/sanitize.ts | Input sanitization |
「wiki-qa」を使用しています。 Where is the API client configured?
期待される結果:
The API client is configured in `src/api/client.ts:22-45`. It uses:
- Base URL from environment variables (line 28)
- Timeout of 30 seconds (line 32)
- Retry logic for failed requests (lines 38-42)
The client is initialized in `src/app.ts:15` and passed to services via dependency injection.
「wiki-qa」を使用しています。 What happens when a user uploads a file?
期待される結果:
File upload follows this workflow:
1. **Request received** - Router routes POST /upload to handler (src/routes/upload.ts:12)
2. **Validation** - File size and type checked (src/middleware/validate.ts:45-52)
3. **Storage** - File saved to temp directory (src/services/storage.ts:78-92)
4. **Processing** - Background job processes file (src/jobs/process.ts:15-30)
Note: Virus scanning is mentioned in comments but not implemented (src/services/storage.ts:85).
セキュリティ監査
安全Static analysis detected 4 potential issues that are all false positives upon manual review. The skill is a legitimate documentation/Q&A tool that reads source code files and answers questions based on file content. No actual code execution, external commands, or cryptographic operations are present.
品質スコア
作れるもの
New developer onboarding
Quickly understand how specific components work by asking natural language questions about the codebase structure and implementation.
Legacy code investigation
Navigate unfamiliar codebases efficiently by asking where features are defined and how different modules interact.
Architecture documentation
Generate explanations of system architecture and component relationships based on actual source code evidence.
これらのプロンプトを試す
How does authentication work in this project?
Where is the user validation logic defined?
How do the frontend and backend communicate in this application?
What could cause the payment processing to fail based on the error handling code?
ベストプラクティス
- Ask specific questions about particular components, functions, or features for the most detailed responses
- Review the Key Files table to understand which files are relevant to your question
- Follow the inline citations to dive deeper into specific implementation details
- Use follow-up questions to explore related components or understand how different parts connect
回避
- Asking questions about topics unrelated to the codebase (this skill only analyzes repository files)
- Expecting runtime behavior analysis without looking at test files or execution traces
- Assuming answers include information from external documentation or APIs
- Requesting code execution or testing functionality (this is read-only analysis)