What problem does it solve?
Preloading every Google product skill into context is expensive and impractical. This Skill fetches a remote catalog index of published Google skills only when needed, matches the request against skill descriptions, and loads just the relevant entry points.
Core Features & Use Cases
- On-demand catalog lookup: Fetches the skills index from the google/skills GitHub repository with byte-exact retrieval and validates the response before use.
- Request routing: Matches user requests against skill descriptions covering Google Cloud, Google Ads, Google Workspace, Android, Firebase, YouTube, Maps, Gemini, and more, shortlisting at most three candidates.
- Failure handling: Provides an ordered fallback path when the catalog fetch fails, including a repository tree listing via the GitHub API, and mandates honest reporting when no skill applies.
- Use Case: A user asks how to deploy a model on Vertex AI. The Skill fetches the catalog, filters entries matching Vertex AI, retrieves the matching skill's entrypoint, and hands off to that skill's instructions.
Quick Start
Find the right Google skill for my request about deploying a container to Cloud Run and load its instructions.