Extract

Extract Codebase Knowledge

Build or rebuild the .gauntlet/knowledge.json knowledge base.

When NOT To Use

• Tribal knowledge no parser can see (use gauntlet:curate)
• Building the code graph (use gauntlet:graph-build)

Steps

1. Identify target directory: use the current working directory or a user-specified path

2. Run AST extraction: invoke the extractor script

   python3 ${CLAUDE_PLUGIN_ROOT}/scripts/extractor.py <target-dir>
   

3. AI enrichment: for each extracted entry, enhance the detail field with natural language explanation of business logic, data flow, architectural role, and rationale

4. Cross-reference: link related entries across modules by matching imports, shared types, and data flow paths

5. Merge with annotations: preserve existing curated entries in .gauntlet/annotations/

6. Save: write to .gauntlet/knowledge.json

7. Report: show summary by category, coverage gaps, difficulty distribution

Exit Criteria

• .gauntlet/knowledge.json exists and is valid JSON after the skill completes; entries from .gauntlet/annotations/ are merged and not overwritten
• Report shows entry counts broken down by all 7 categories (business_logic, architecture, data_flow, api_contract, pattern, dependency, error_handling) with coverage gaps identified
• Each extracted entry has a detail field containing a natural language explanation (not just the raw AST node name)
• Cross-reference links between related entries are present for modules sharing imports, shared types, or data flow paths

Category Priority

1. business_logic (weight 7)
2. architecture (weight 6)
3. data_flow (weight 5)
4. api_contract (weight 4)
5. pattern (weight 3)
6. dependency (weight 2)
7. error_handling (weight 1)

Source: athola/claude-night-market → plugins/gauntlet/skills/extract/SKILL.md