Skilled

TL;DR

A dashboard that identifies which AI agent functions are just wasting context space and burning tokens without providing value.

Who is this actually for?

Engineers managing complex multi-agent workflows who have reached the 'spaghetti code' phase of tool integration.

The Good

• Reduces prompt bloat by identifying and removing unused tool definitions.
• Provides a clear map of agent capabilities to help debug logic overlaps.

The Catch (Potential Downsides)

Integration likely requires piping your execution logs to yet another third-party service. It also risks flagging 'rare but critical' safety skills as redundant.