sickn33/agentic-awesome-skills

Agentic Awesome Skills frames itself as a control plane that lets agents discover and select from a catalog of nearly 2,000 pre-built skills rather than having developers wire every capability by hand. The included stack — CLI, local MCP, plugin system, and a workbench UI — covers the tooling layer a team would otherwise assemble from parts. The catalog-first model is the distinguishing angle: instead of coding a skill per task, the idea is that the agent figures out what it needs from what is available. That is genuinely interesting as a design philosophy. The reservation is that the repo sits at an early stage and a skill catalog of nearly 2,000 entries is only useful if the quality bar on individual skills is consistent — which is hard to verify without running a lot of them. Worth a look for AI engineers building complex agent pipelines who want a library-first starting point rather than a blank orchestrator. -> Best for: AI engineer prototyping a multi-skill agent pipeline who prefers a curated catalog over building from scratch.