What problem does it solve?
AI agents often need reusable registry-based skills but browsing prompts.chat, reading metadata, and installing files manually is fragmented and error-prone, wasting time for builders. This workflow unifies MCP search, metadata review, retrieval, and installation so you can bring new capabilities online with confidence.
Core Features & Use Cases
- Catalog Search & Discovery: Query prompts.chat via search_skills using keywords, categories, and tags to surface the best matches quickly.
- Metadata & Result Presentation: Summarize titles, descriptions, authors, file lists, categories, and tags so the user can compare options before installing.
- Retrieval & Installation Workflow: Use get_skill to download every file, write them to .claude/skills/{slug}/, and double-check that SKILL.md retains the required frontmatter before confirming activation.
- Use Case: When a user asks for a code-review assistant, search for the appropriate skill, show the metadata, install the files, and explain when it will trigger inside the AI assistant.
Quick Start
Call search_skills with the user's keywords, pick the best match, run get_skill, and save every returned file under .claude/skills/{slug}/ while confirming SKILL.md still contains the required frontmatter.