Analyze codebase structure, generate domain-specific sub-skills (UI, Backend, Database, etc.), and create agent-guidance files to help AI agents navigate and develop consistently within a project. Use when onboarding to a new codebase, creating project documentation, or setting up agent guidance systems.
A skill for analyzing existing codebases and creating structured guidance systems for AI agents. Generates domain-specific knowledge and navigation aids without modifying source code locations.
Use this skill when:
python ./skills/project-index/scripts/scan_codebase.py /path/to/project --output codebase_index.json
Run the scanning script to generate a complete structural map:
python ./skills/project-index/scripts/scan_codebase.py <project_path> --output codebase_index.json
Important: The script provides a structural overview. You MUST also:
Parameters:
<project_path>: Absolute path to the project root--output: Output filename (default: codebase_index.json)Excluded directories: .git, node_modules, __pycache__, dist, build, .next, .expo
Review the generated codebase_index.json and:
CRITICAL: Do not rely solely on the JSON index. Actually explore the codebase:
For each identified domain, create a focused sub-skill using skill-writer:
# Load the skill-writer skill first, then:
# Create sub-skills like project-index-ui, project-index-backend, etc.
Each sub-skill should:
Use the template: ./skills/project-index/templates/skill-template.md
Naming convention: project-index-<domain> (e.g., project-index-ui, project-index-backend)
Create agent-<name>.md files in relevant project directories:
Use the template: ./skills/project-index/templates/agent-template.md
Each file should contain:
Location: ./skills/project-index/templates/skill-template.md
---