rules-distill

To extract common rules from multiple LLM skills, run a distillation workflow that inventories installed skills, clusters recurring cross-cutting themes, and drafts reusable governance rules for standardizing behavior across projects.

Cross-cutting pattern extraction for skill governance is the process of cross-reading installed skills to identify consistent principles, then consolidating them into versioned rule files with provenance tracking for audit purposes.

You generate rule drafts from existing skills by cross-reading the skill inventory against current rule files to produce proposed appends or revisions, storing results in a structured JSON file, and requiring explicit user approval before applying any changes.

Yes, you need jq installed to run the skill distillation workflow, as the underlying scripts component relies on this dependency to process structured JSON data and manage rule extraction outputs.

The best way to consolidate recurring principles from an LLM skill inventory is to use an automated pattern-extraction tool that clusters themes across the full inventory and cross-reads them against existing rules to surface consistent governance updates.

You should avoid automating rule extraction from your skills when you lack existing rule files for cross-reading or when your project requires manual review of every principle, as the workflow drafts changes that still require explicit user approval before applying.