Find Skill

**Judgement:** Strongly recommended as a meta-skill for expanding an agent’s capabilities via the open skills ecosystem, especially when users ask for specialized or recurring tasks. **What it does:** - Listens for intent like “how do I do X”, “find a skill for X”, or “is there a skill that can…”. - Uses `npx skills find [query]` to search for relevant skills on skills.sh. - Presents candidate skills with descriptions, install commands, and links. - Can install chosen skills with `npx skills add <owner/repo@skill> -g -y`. - Falls back to built-in AI help if no skills match and can suggest creating a new skill. **Key strengths:** - Well-aligned with the Skills ecosystem and the Skills CLI; clear, step-by-step operational guidance. - Very high usability—simple trigger conditions and example-driven instructions. - Encourages structured discovery across common domains (web dev, testing, DevOps, docs, design, productivity). **Main risks / limitations:** - **Third‑party skill trust:** It may recommend skills from external repos; actual safety and quality depend on those packages (supply-chain risk, unvetted code or tools). - **Dependency on external infrastructure:** Functionality relies on the Skills CLI, network connectivity, and the skills.sh index; outages or API/CLI changes can reduce reliability. - **Result quality:** Search quality is only as good as the underlying registry and queries; may miss niche or poorly tagged skills. **Recommended scenarios:** - Agents embedded in developer, DevOps, or content workflows where users often ask for “a better way” or “a tool” to handle tasks. - Systems that want to stay lean by default but dynamically extend capabilities only when the user’s needs justify installing a dedicated skill. - Environments where operators are comfortable reviewing and trusting third-party skills discovered via the Skills CLI before heavy use.