Scout
Map unfamiliar codebases in minutes using parallel agents that explore structure, patterns, and conventions simultaneously.
What This Skill Does
The Challenge: Working on large codebases is slow without orientation. Developers and AI agents waste time searching for the right files, re-discovering patterns, or writing code that conflicts with existing conventions.
The Solution: Scout skill deploys parallel agents to explore codebase structure, identify key files, map architectural patterns, surface naming conventions, and produce an orientation report — all before a single line of code is written.
Activation
Implicit: Activates at the start of implementation tasks in unfamiliar repositories.
Explicit: Activate via prompt:
Activate scout skill to map the authentication system in this codebase
Capabilities
1. Structural Mapping
Identify directory structure, module boundaries, and file organization patterns.
Mapped elements:
- Entry points (
main.ts,index.ts,app.ts) - Configuration files (
.env.example,tsconfig.json, framework configs) - Key directories (
src/,lib/,components/,api/,services/) - Test organization (
__tests__/,*.spec.ts,*.test.ts)
2. Pattern Recognition
Surface recurring patterns, conventions, and architectural choices.
Patterns analyzed:
- State management (Redux, Zustand, Context API, Jotai)
- API patterns (REST, GraphQL, tRPC, server actions)
- Styling approach (CSS modules, Tailwind, styled-components)
- Error handling conventions
- Authentication patterns
3. Dependency Analysis
Map external dependencies and their usage patterns.
# Key files to scan
package.json # Dependencies and scripts
tsconfig.json # TypeScript config and paths
.env.example # Required environment variables
4. Parallel Exploration
Deploy multiple agents to explore different subsystems simultaneously.
Parallel agent assignments:
- Agent A: Database models and schema
- Agent B: API routes and controllers
- Agent C: Frontend components and state
- Agent D: Configuration and infrastructure
Prerequisites
- Repository access (local or via Repomix)
- Target area of exploration defined (full codebase or specific feature)
Configuration
No configuration required. Scout operates through file system exploration.
Scope options:
- Full repository scan (for new codebases)
- Feature-scoped scan (for targeted implementation)
- Dependency scan (for integration work)
Best Practices
1. Scout before you plan Run scout before creating implementation plans. Architecture findings often change the approach.
2. Define the question before scouting “Where is authentication handled?” produces better scout output than “tell me about the codebase.”
3. Save scout reports
Scout outputs are valuable orientation documents. Save to plans/reports/scout-report.md.
Common Use Cases
Use Case 1: Onboarding to New Codebase
Scenario: Developer joins team and needs orientation on large Next.js app.
Workflow:
- Run scout on full repository
- Map: page structure, API routes, data layer, shared components
- Identify: key patterns, naming conventions, testing approach
- Generate orientation report
Output: Scout report with architecture overview, key file list, and convention guide.
Use Case 2: Pre-Implementation Research
Scenario: Adding payments feature to existing app. Need to understand current auth, user model, and API patterns.
Workflow:
- Scout scope:
src/api/**,src/models/**,src/lib/** - Map: existing payment-adjacent code, user model fields, API auth pattern
- Identify: where to add routes, which patterns to follow
- Feed findings to planner
Output: Implementation context report fed into planning phase.
Troubleshooting
Issue: Scout report is too high-level, missing specific patterns Solution: Narrow scope with specific directories or questions.
Issue: Large codebase scout takes too long Solution: Use Repomix to pack codebase first, then scout the packed file.
Related Skills
- Repomix - Pack repo for full-context analysis
- Plan - Create implementation plan from scout findings
- Research - Deep technical research
Related Commands
/ckm:scout- Codebase scouting workflow/ckm:repomix- Repository packing for analysis/ckm:plan- Implementation planning