| name | codebase-scan |
| description | Create a high-level map of the project structure and identify architectural pillars, entry points, and module boundaries. Activate at Bootstrap initialization or before architecture extraction. |
| license | MIT |
| compatibility | Works with any filesystem-based AI coding agent |
| metadata | {"author":"gaai-framework","version":"1.0","category":"cross","track":"cross-cutting","id":"SKILL-CODEBASE-SCAN-001","updated_at":"2026-02-26T00:00:00.000Z","status":"stable"} |
| inputs | ["repository/**"] |
| outputs | ["codebase_tree","key_files_list"] |
Codebase Scan
Purpose / When to Activate
Activate:
- During Bootstrap initialization (before architecture extraction)
- When onboarding GAAI onto an existing codebase
- When the project structure is unknown
Process
- Recursively list directories and files
- Identify entry points, config files, manifests, and core modules
- Highlight likely architecture boundaries
- Output a structured project tree and key file list
Outputs
codebase_tree — structured directory tree with annotations for key files
key_files_list — list of significant files: entry points, configs, manifests, core modules, test runners
Quality Checks
- Repository structure is fully visible
- Core technical pillars are identified
- No blind spots remain
- Output is structured for use by
architecture-extract
Non-Goals
This skill must NOT:
- Interpret architecture decisions (use
architecture-extract)
- Read file contents in depth
- Make recommendations
Maps the terrain. Does not interpret it.