AAS Core Ships a Local Control Plane That Discovers, Validates and Plans Exact Skill Stacks From 1,965+ Agentic Skills

GitHub·high signal

sickn33/agentic-awesome-skills (43,574 stars) is not another awesome-list — it is a Python control plane with a CLI and local MCP server that recommends and validates which skills an agent should load for a given task, backed by a catalog of 1,965+ skills. This targets the skill-sprawl problem that arrives once a team has dozens of skill files and no principled way to decide which ones belong in context. Skill selection as a solved, validated step is a meaningfully new layer in the agent stack.