Skill Manager

kkkkhazix's avatarfrom kkkkhazix
⭐0stars🔀0forks📁View on GitHub
[Developer Tools]

When & Why to Use This Skill

The Skill Lifecycle Manager is a specialized utility designed to maintain and optimize a library of GitHub-wrapped Claude skills. It automates the detection of updates, manages local inventories, and streamlines the refactoring process for skill wrappers, ensuring your AI agent's capabilities remain current, secure, and fully functional with the latest upstream changes.

Use Cases

  • Automated Update Monitoring: Periodically scan local skill directories to identify 'stale' tools by comparing local commit hashes against the latest remote HEAD on GitHub.
  • Guided Skill Refactoring: Assist the agent in upgrading skill wrappers by fetching new documentation, performing diff analysis on README files, and updating metadata and wrapper logic to match new CLI arguments.
  • Skill Inventory Management: Maintain a clean development environment by listing all installed skills with their versions and performing permanent deletions of unused or deprecated tools.
nameskill-manager
descriptionLifecycle manager for GitHub-based skills. Use this to batch scan your skills directory, check for updates on GitHub, and perform guided upgrades of your skill wrappers.
licenseMIT

Skill Lifecycle Manager

This skill helps you maintain your library of GitHub-wrapped skills by automating the detection of updates and assisting in the refactoring process.

Core Capabilities

  1. Audit: Scans your local skills folder for skills with github_url metadata.
  2. Check: Queries GitHub (via git ls-remote) to compare local commit hashes against the latest remote HEAD.
  3. Report: Generates a status report identifying which skills are "Stale" or "Current".
  4. Update Workflow: Provides a structured process for the Agent to upgrade a skill.
  5. Inventory Management: Lists all local skills and provides deletion capabilities.

Usage

Trigger: /skill-manager check or "Scan my skills for updates" Trigger: /skill-manager list or "List my skills" Trigger: /skill-manager delete <skill_name> or "Delete skill "

Workflow 1: Check for Updates

  1. Run Scanner: The agent runs scripts/scan_and_check.py to analyze all skills.
  2. Review Report: The script outputs a JSON summary. The Agent presents this to the user.
    • Example: "Found 3 outdated skills: yt-dlp (behind 50 commits), ffmpeg-tool (behind 2 commits)..."

Workflow 2: Update a Skill

Trigger: "Update [Skill Name]" (after a check)

  1. Fetch New Context: The agent fetches the new README from the remote repo.
  2. Diff Analysis:
    • The agent compares the new README with the old SKILL.md.
    • Identifies new features, deprecated flags, or usage changes.
  3. Refactor:
    • The agent rewrites SKILL.md to reflect the new capabilities.
    • The agent updates the github_hash in the frontmatter.
    • The agent (optionally) attempts to update the wrapper.py if CLI args have changed.
  4. Verify: Runs a quick validation (if available).

Scripts

  • scripts/scan_and_check.py: The workhorse. Scans directories, parses Frontmatter, fetches remote tags, returns status.
  • scripts/update_helper.py: (Optional) Helper to backup files before update.
  • scripts/list_skills.py: Lists all installed skills with type and version.
  • scripts/delete_skill.py: Permanently removes a skill folder.

Metadata Requirements

This manager relies on the github-to-skills metadata standard:

  • github_url: Source of truth.
  • github_hash: State of truth.