claude skill auditor fixer

The problem

Skills often fail in production due to structural drift, hallucination traps, and triggers that are either too broad or self-referential. Manually auditing SKILL.md files for marketplace compliance is tedious and prone to missing subtle YAML or logic errors.

What it does

• Performs static analysis on SKILL.md files via file path, skill name, or raw text input.
• Identifies critical defects like self-validation traps where a skill incorrectly triggers itself.
• Validates YAML frontmatter, trigger specificity, and line count constraints.
• Detects scope creep where the description promises features absent from the body.
• Generates a repaired SKILL.md file alongside a severity-ranked findings report.

Why this beats prompting it yourself

General-purpose prompts often miss strict platform constraints like trigger count limits or specific YAML formatting rules. This skill uses a structured checklist to ensure marketplace readiness without the risk of the auditor introducing its own hallucinations into your documentation.

Use cases

• Verify production-readiness before submitting a skill to the Agensi marketplace.
• Debug why a skill is triggering on unrelated prompts or failing to invoke.
• Standardize a library of legacy skills to meet current platform specifications.
• Generate automated audit reports for CI/CD skill deployment pipelines.

Known limitations

Static analysis only. It does not execute the skill to test runtime behavior or external tool availability. Markdown rendering is not validated.