SkillRanker
Rust CLI that ranks which agent skills fit the next step from live session context with Jev, can abstain, and hooks into Claude Code.
Overview
SkillRanker (sr) is a standalone Rust CLI that helps agents pick the right skill for the next step. It combines the recent conversation, current request, workspace signals and the harness's visible skill inventory, then uses a Jev-compatible provider in two passes: a broad comparison of candidates, then a closer evaluation of a shortlist using richer excerpts. Both passes include a real 'none of these' option, and the result is advisory. For libraries over 254 eligible skills, a local Quill search from FrankenSearch narrows candidates first; explicit skill requests resolve locally. It builds on TypeSafe's skill-suggestion cookbook.
It integrates with Claude Code through a UserPromptSubmit hook, outputs structured JSON, and includes --why-not tracing of excluded candidates, opt-in case capture with offline replay, local feedback and calibration, evaluation and risk monitoring, and an inline TUI. TypeSafe is the default provider, with Cloudflare-hosted Jev as an alternative; there is no local model, so fresh rankings need your own credential and network access.
What the decision model decides
| Type | Question | Answers |
|---|---|---|
| Choice | Which of these skills (or none) fits the next step? | — |
| Yes / no | Does this shortlisted skill fit the next step? | — |
How it works
- Collect session context and the visible skill roster
- Narrow large rosters locally with Quill
- Jev compares candidates, then evaluates a shortlist with a none option
- Advise the agent via hook or JSON; record feedback
Strengths
- Can abstain rather than force a skill
- Explain and replay results without new calls
- Local feedback for calibration
Caveats
- Requires your own provider credential
- Custom licence
Getting started
sr demo --case useful to inspect an offline fixture, then set TYPESAFE_API_KEY and run sr rank --allow-network.At a glance
- Type
- CLI
- Status
- beta
- Language
- Rust
- Runtime
- CLI + Claude Code hooks
- Has code
- yes
- License
- Custom license
- Stars
- 134
Use cases and integrations
How curated lists describe it
Coding agents: standalone Rust CLI that uses Jev to rank candidate skills against live session context, advising the next step through a Claude Code UserPromptSubmit hook.
yibie/awesome-jevRust CLI: Jev ranks which agent skill fits the next step from live session context, with Claude Code hooks
AnotiaWang/awesome-decision-modelsA standalone Rust CLI powered by TypeSafe Jev that ranks agent skills for the next step using live session context, with local safeguards and feedback.
logicrw/awesome-jev-projectsRust CLI powered by Jev from TypeSafe.ai that ranks agent skills for the next step using live session context. Includes Claude Code hooks, structured JSON, abstention, and local feedback. Requires a TypeSafe API key.
wh000wh000/awesome-jev-liveRust CLI that ranks agent skills against live session context and can abstain.
cobanov/awesome-jevRust CLI powered by Jev from TypeSafe.ai that ranks agent skills for the next step using live session context. Includes Claude Code hooks, structured JSON, abstention, and local feedback. Requires a TypeSafe API key.
hellogumbo/awesome-jevRust CLI that ranks which agent skills fit the next step from live session context, with Claude Code hooks.
MrJev/awesome-jev
Frequently asked questions
What does Jev decide in SkillRanker?
Which skills, if any, fit the agent's next step, in a broad pass and a shortlist pass.
Can it say no skill fits?
Yes, both passes include a 'none of these' option.
Does it include a local model?
No; it needs a Jev-compatible provider such as TypeSafe or Cloudflare.
Related projects
Claude Code plugin that sends judgment calls to Jev, acts at 0.8+ for yes or 0.2- for no, and gathers more evidence in between.
Agent skillbetaJavaScriptOpen-source Jev skills for Claude Code, Codex, Cursor and Grok: one client to gate, route, guard, verify, triage, review and act.
ProjectbetaTypeScriptLocal gateway and MCP tools that let Jev predict which tool a coding agent should call next, from suggestions to enforced choices.
MCP serverbetaTypeScriptFourteen opt-in Claude Code and Codex skills where Jev makes one bounded choice: which skill, model, test, file or next step.
ProjectbetaPythonToolkit where Jev condenses large MCP responses for Claude Code and maps Codex requests to reviewed, fixed repository operations.
ProjectbetaTypeScriptClaude Code function hook that sends the agent's this-or-that questions to Jev and optionally ranks installed skills per prompt.
ProjectbetaTypeScript