rules-distill

To extract cross-cutting rules from installed skills, run an automated distillation workflow that scans your inventory, identifies overlapping principles, and generates formal rule candidates for review.

The best way to identify overlapping principles is through automated inventory and LLM-based judgment, which scans installed skills and rules to surface cross-cutting patterns and clusters.

LLM judgment for rules distillation works by applying a formal verdict workflow across your skills inventory, determining whether to Append, Revise, create a New Section, create a New File, or mark as Already Covered.

Yes, you can see rule references across all your installed skills to ensure traceability, allowing you to track where specific principles and cross-cutting rules are applied.

You need to run a rules distillation workflow for process improvement when your installed skills accumulate overlapping principles that require formal consolidation into a structured rules corpus.

A key limitation of automated rules distillation is that LLM-assisted verdicts only generate rule candidates for review, meaning human validation is still required before applying changes to your skills inventory.