What problem does it solve?
When you maintain a library of custom skills for your team, common principles and guardrails are often scattered across multiple skill files, leading to inconsistent practices, duplicated effort, and missing standardized rules that should apply across all team workflows.
Core Features & Use Cases
- Exhaustive Skill & Rule Inventory: Uses deterministic shell scripts to scan all installed skills and existing rule files, capturing full frontmatter, content, and rule headings with no missed entries.
- Contextual Principle Extraction: Leverages LLM judgment to cross-read all skill and rule content, identifying actionable cross-cutting principles that appear in 2 or more skills, have clear violation risks, and are not already covered by existing rules.
- Structured Rule Recommendations: Generates detailed verdicts for each candidate principle, including whether to append to existing rules, revise outdated content, add new sections, or create new rule files, with draft text and evidence links for transparency.
- Use Case: For a software team using 12 custom skills for coding, security, and agent testing, this skill can automatically surface the shared principle "sanitize all LLM output before storing or re-injecting it" that appears in 3 separate skills, and propose adding it to the team's shared security rule file, cutting hours of manual cross-skill review.
Quick Start
Use the rules-distill skill to scan your installed skills and generate a structured report of candidate principles to add to your team's shared rule files.
