Explore codebase, understand architecture, find how things work. Navigate unfamiliar repositories, discover code structure, find where functionality lives by intent rather than exact text. NOT for debugging failures (use debugging) or checking change impact (use change-planning).
Semantic code exploration that finds execution flows by intent, not text. Use when navigating unfamiliar repos to discover how features work by reading map://current overview, drilling into communities, and searching for concepts like "payment validation" to get ranked processes with file paths.
/plugin marketplace add noodlbox/noodlbox-claude-plugins/plugin install noodlbox@noodlboxThis skill inherits all available tools. When active, it can use any tool Claude has access to.
Semantic code exploration that returns processes (execution flows) instead of isolated file matches.
1. READ map://current → Get codebase overview (~200 tokens)
2. READ map://current/community/{id} → Drill into relevant module
3. noodlbox_query_with_context → Search for specific concepts
Copy and track progress:
Exploration Progress:
- [ ] Read map overview
- [ ] Identify relevant community
- [ ] Drill into community detail
- [ ] Search for specific symbols
- [ ] Read source files
Semantic search returning processes ranked by relevance.
noodlbox_query_with_context(
repository: "current",
q: "payment validation",
task_context: "exploring payment system",
current_goal: "understand validation flow",
search_intention: "find entry points",
limit: 5,
max_symbols: 10
)
Output structure:
Process: "Payment validation starting with validatePayment"
├── validatePayment (src/payments/validator.ts:42) ← matched
├── checkAmount (src/payments/validator.ts:78)
├── verifyCard (src/payments/card.ts:15)
└── processTransaction (src/payments/processor.ts:23)
Start with limit: 3-5. Increase only if needed.
Codebase overview with communities, stats, and cross-flows.
repository: my-project
stats: { communities: 15, symbols: 2400, processes: 890 }
communities:
- id: abc123, label: PaymentProcessing, symbols: 47
- id: def456, label: UserAuthentication, symbols: 32
cross_flows:
- from: PaymentProcessing, to: UserAuthentication, calls: 12
Token cost: ~200 tokens. Read first when unfamiliar with codebase.
Community detail with symbols ranked by importance.
id: abc123
label: PaymentProcessing
symbols:
- name: validatePayment, centrality: 0.92, file: src/payments/validator.ts
- name: processCharge, centrality: 0.78, file: src/payments/processor.ts
entry_points:
- name: handlePaymentRequest, callers: [APIGateway, WebhookHandler]
processes:
- id: xyz789, label: "Payment validation flow"
Token cost: ~500 tokens. Use after identifying relevant community.
Full execution trace with file paths and line numbers.
Use when: Need complete call chain for a specific flow.
| Term | Meaning |
|---|---|
| Community | Cluster of tightly-coupled symbols (functional module) |
| Process | Execution flow from entry point through call chain |
| Centrality | Symbol importance - high means many callers/callees |
| Entry point | Symbol called from outside its community |
Task: "How does authentication work in this codebase?"
Step 1: Get overview
READ map://current
→ See "UserAuthentication" community (32 symbols, high cohesion)
Step 2: Drill into auth community
READ map://current/community/{auth-id}
→ Key symbols: login, validateToken, refreshSession
→ Entry points: handleLoginRequest, authMiddleware
→ Process: "Login flow starting with validateCredentials"
Step 3: Search for specifics
noodlbox_query_with_context(q: "token validation", limit: 3)
→ Focused results from the auth area
→ Each symbol has file_path and line numbers
Step 4: Read source
Read src/auth/token.ts
Checklist for this example:
- [x] Read map overview
- [x] Identify relevant community (UserAuthentication)
- [x] Drill into community detail
- [x] Search for specific symbols (token validation)
- [x] Read source files (token.ts)
| Need | Use Instead |
|---|---|
| Check change impact | change-planning skill |
| Debug failing code | debugging skill |
| Plan refactoring | refactoring skill |
| Find exact string | Grep (faster for literal matches) |
| Read specific file | Read tool directly |