Agent Skills Search

**Judgement:** A focused, useful search skill for discovering tools, skills, and MCP servers within the Agnxi.com ecosystem. It excels as a discovery layer rather than a full information source. **What it does well:** - Performs keyword-based search over a curated directory of AI agent tools, MCP servers, and skills. - Returns direct URLs, making it easy to chain with browser/GitHub capabilities for deeper inspection. - Simple interface (one required `query` parameter) and no API keys required. **Key risks / limitations:** - Scope is limited to what Agnxi indexes; it will miss tools/skills not present in that directory. - Results may include outdated or broken links if the underlying sitemap is stale or sites have changed. - Depends on a local Python script and external website availability; failures or latency are possible and should be handled gracefully. - Returned URLs are not validated semantically; agents must not assume relevance without checking content. **Recommended scenarios:** - When an agent needs to **discover new skills or MCP servers** to solve a task (e.g., “browser automation”, “Postgres MCP”). - As a first step in **tool selection workflows**, followed by using browser/GitHub tools to vet documentation, licenses, and maintenance state. - For **exploratory research** across AI-agent-related tooling, where breadth of options within the Agnxi ecosystem is more important than deep detail from a single source.