What problem does it solve?
Teams maintaining a library of AI skills need a consistent way to check whether a skill meets repository standards before shipping or after edits. This Skill audits an existing pm-skills skill against structural conventions and quality criteria, producing a structured validation report with severity-graded findings and actionable recommendations.
Core Features & Use Cases
- Tier 1 Structural Checks: Deterministic checks mirroring CI, covering frontmatter fields, naming, file presence, description length, and eval-asset contracts.
- Tier 2 Quality Checks: LLM-assessed coherence checks such as template coverage, checklist testability, example completeness, and placeholder leakage.
- Machine-Readable Report: Emits a pipe-delimited report (schema v1) that humans can scan and that the iterate skill can consume as input.
- Batch Mode: Run structural checks across all skills with
--all to audit the catalog after a convention change.
- Use Case: After editing a skill's SKILL.md, ask for a validation report to confirm it still passes every structural and quality check before opening a pull request.
Quick Start
Validate the deliver-prd skill and give me a full pass/fail report with severity-graded findings before we ship it.
