Medium LLM
Everyone Is Writing Skills for Their Agents. Almost Nobody Can Say When They’re Complete.
agenticMCP
What happened
A brief post discussing the rise of 'agent skills' (which Simon Willison called 'a bigger deal than MCP') and the challenge of defining when a skill file is complete and contains sufficient expert knowledge.
Why it matters
Defining completeness and evaluation metrics for agent skills is an open challenge in agentic design.
The take
The debate between standardized protocols (MCP) and custom 'agent skills' is highly relevant. However, because this is a thin teaser, it raises the right question ('when is a skill complete?') without providing the architectural answers.
Do this
Monitor the discussion around 'agent skills' as an alternative or complement to MCP-based tool definitions.
Don't read this site daily. Get it in your inbox.
The daily brief and Sunday deep dive — distilled, scored, and opinionated. For builders only.