Overview & result
SkillRanker is a Rust CLI that uses Jev to rank available agent skills against the live session context. It adds local safeguards, feedback, and replay around the ranking. The repository does not support the previous sub-millisecond overhead claim.
From the original post
The right skill for the next step, powered by Jev from TypeSafe.ai.
A standalone Rust CLI that puts TypeSafe.ai's Jev at the center of skill selection: Jev evaluates your agent's live context, compares the available skills, and estimates which ones fit the next step. SkillRanker supplies the session integration, local safeguards, and inspectable feedback around it.
How Jev fits in the loop
- Parse installed skill metadata directory.
- Dispatch query to Jev for fast ranking.
- Emit ranked skills to stdout for shell scripts or agent harnesses.
How to reproduce
- Inspect the source implementation at https://github.com/Dicklesworthstone/skillranker.
- Configure local Jev API credentials and register required tool/model bindings.
- Execute test prompts and observe decision latency and routing outputs.
Why this build matters
Provides a reproducible example of Jev making a bounded skill-selection decision while the surrounding agent handles execution and feedback.
Limitations
- It includes no local model: fresh ranking needs your own provider credential and network access.
- It recommends skills only; it does not execute them, grant permissions or override the agent's instructions.
- Installation requires Bash, Python 3 and curl; source builds also need Git and Rust.