SkillRanker
A Rust CLI that uses TypeSafe.ai's Jev decision model to rank which agent skill best fits the next step of a coding session, with hooks, JSON output, and local feedback.
SkillRanker is a standalone Rust command-line tool that ranks an agent's available skills for its next step. It gathers the session's recent conversation, current request, and workspace signals, then sends the candidates to Jev, a decision model from TypeSafe.ai, which compares them and can also answer that none of them fit.
For libraries with many eligible skills, a local component narrows the candidates before Jev evaluates them, while smaller rosters go to Jev in full, and an explicit skill request resolves locally without either stage. Results are advisory: the agent still decides what to consult.
SkillRanker includes hooks for Claude Code, structured JSON output, an offline demo fixture, and local feedback and calibration, but it has no local model of its own and requires a TypeSafe API key to rank skills.