The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses Bash to execute internal Python scripts (scripts/gap-analyzer.py, scripts/validate-decomposition.py, scripts/audit-reference-depth.py). These are used for structural validation and domain analysis of reference files.
[DYNAMIC_EXECUTION]: The skill generates reference files based on templates and interpolates research findings into these files. It uses a "Headless Prompt" workflow where the AI acts as an autonomous process to commit and push changes via GitHub CLI (gh).
[SAFE_PRACTICE]: The skill includes extensive safety gates, mandatory validation steps, and snapshot/restore mechanisms (/tmp/decomp-before-{name}.md) to prevent data loss during content decomposition.
[SAFE_PRACTICE]: The gap-analyzer.py script uses deterministic regex patterns to identify missing sub-domains, and the quality-rubric.md provides strict criteria to prevent generic "Level 1" (thin) reference bloat.

reference-enrichment