Can it detect infinite loops in skill calls?

Yes, it traces all Skill() and Agent() references to build a directed graph, identifying both direct (A to B, B to A) and indirect circular dependencies.

What is the difference between this and a standard linter?

While structural linters check if files exist and follow schema, the Skill Lifecycle Auditor performs semantic analysis to ensure that the logic, instructions, and dependencies between skills actually make sense for the AI's execution.

Where does the skill save its findings?

The skill generates comprehensive audit reports, Mermaid-format call graphs, and prioritized recommendations in the .claude/audits/ directory of your project.

Does it automatically fix the issues it finds?

No, this is a diagnostic tool designed to provide deep visibility. It generates a detailed recommendations.md file with prioritized fixes for the developer to implement.

Skill Lifecycle Auditor

Name: Skill Lifecycle Auditor
Author: Jamie-BitFlight

byJamie-BitFlight

•

Security & Testing

Performs deep semantic validation of skill interconnections, dependency chains, and instruction coherence within Claude plugins.

The Skill Lifecycle Auditor provides a comprehensive semantic quality gate for developers building complex Claude plugins. It goes beyond simple structural checks to trace call chains, detect circular dependencies, and identify contradictory instructions that can cause AI workflows to fail. By analyzing the bidirectional coherence of your skills and identifying redundant datasets, this tool ensures your plugin forms a logical, non-contradictory, and efficient workflow. It generates detailed reports, visual call graphs, and actionable recommendations to the .claude/audits/ directory, making it an essential utility for debugging and marketplace preparation.

Key Features

01Detects direct and indirect circular loading patterns to prevent infinite recursion.

02Traces call chain completeness to ensure every skill has the necessary supporting capabilities.

03Discovers manual command sequences that are candidates for script-based automation.

04Scans for duplicated datasets and content blocks for shared reference optimization.

05Identifies unguarded instruction contradictions across related domain skills.

0626 GitHub stars

Use Cases

01Marketplace submission preparation to ensure high semantic quality and plugin stability.

02Debugging complex multi-skill workflows where agents fail to achieve stated goals.

03Refactoring large skill repositories by identifying semantic gaps and redundant instructions.

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add jamie-bitflight/claude_skills audit-skill-lifecycle

For use in Claude.ai and ChatGPT

Download Skill