skill-sync_skill

Overview

This skill keeps Auto-invoke sections in AGENTS.md files synchronized with skill metadata. It detects changes to a skill's scope or auto-invoke settings and regenerates the Auto-invoke tables for the relevant AGENTS.md files. Use it after creating or updating a skill to ensure documentation and workflows stay aligned.

How this skill works

The sync scans all skill metadata files to collect scope and auto-invoke entries. It builds or updates the "Auto-invoke Skills" table for each AGENTS.md corresponding to the declared scopes. The script can run for all scopes, a single scope, or as a dry run to preview changes before applying them.

When to use it

• After adding a new skill that should be auto-invoked
• When changing a skill's metadata.scope or metadata.auto_invoke
• Before releasing or merging changes that affect agent workflows
• When regenerating documentation to reflect current skill behavior

Best practices

• Always include metadata.scope and metadata.auto_invoke in skill metadata for auto-invoke candidates
• Use concise, action-oriented descriptions for auto_invoke entries
• Run the sync script as part of your local validation or CI checks
• Use --dry-run first to review changes before committing updates

Example use cases

• Add a UI skill with scope set to ui and an action like "Creating/modifying React components" so the ui agent lists it in Auto-invoke
• Update a skill that now applies to both ui and api scopes; run sync to update both AGENTS.md files
• Refactor a component and add an auto_invoke entry; run sync with --dry-run to preview documentation edits
• Run the script in CI to ensure Auto-invoke tables are always current before merging