jev-skills
Claude Code and Codex skills powered by Jev, published with a calibration curve measured on 4,995 real agent decisions.
This summary is based on how curated lists describe the project; its README was not available to review.
Overview
This repository provides Claude Code and Codex skills that use Jev, TypeSafe's System One model, for agent judgments. Directory listings describe it as including what it calls the first published Jev calibration curve, measured on 4,995 real agent decisions.
The listing reports 256 calibrated judgements for $0.0005 in 0.72 seconds and frames this as 360 times cheaper than Claude Opus 5. Another listing describes the skills as helping shortlist files, triage findings and extract relevant log lines, and notes no repository licence at the time of review. No README was available for this entry, so details beyond these summaries are unknown.
What the decision model decides
| Type | Question | Answers |
|---|---|---|
| Yes / no | Agent judgment calls (e.g. is this file or finding relevant) answered as calibrated probabilities(inferred) | — |
How it works
- Agent skill sends a batch of judgments to Jev
- Jev returns calibrated probabilities
- Agent acts on results above a threshold
Strengths
- Listing reports 256 judgments for $0.0005 in 0.72 s
- Publishes a calibration curve on 4,995 real agent decisions per the listing
Caveats
- No README available to verify details
- Listed as having no repository licence at review
At a glance
- Type
- Project
- Runtime
- Claude Code / Codex skills
Use cases and integrations
How curated lists describe it
Claude Code & Codex skills powered by Jev, TypeSafe's System One model. 256 calibrated judgements for $0.0005 in 0.72s — 360x cheaper than Claude Opus 5. Includes the first published Jev calibration curve, measured on 4,995 real agent decisions.
logicrw/awesome-jev-projects
Frequently asked questions
What is notable about this repo?
Listings say it includes a Jev calibration curve measured on 4,995 real agent decisions.
Which agents does it support?
Claude Code and Codex, according to the listing.
Is it licensed?
A listing notes it had no repository licence at review.
Related projects
Agent skill that tells you when you need an eval and how to build it, with an opt-in lane routing closed-label residue to Jev.
BenchmarkbetaPythonEval-first MCP server for Jev: prototype questions, map them over many items, and measure accuracy and thresholds on labelled data.
MCP serverbetaTypeScriptAgent skills for placing, building, evaluating and improving decision-model systems, with Jev as the default hosted example.
Agent skillbetaMCP server exposing Jev as one evaluate tool, plus a harness comparing Jev's judgments with general LLMs on shared datasets.
MCP serverbetaJavaScriptClaude Code and Codex skills plus a study course for building with Jev, including Jevaluate, which rates how well projects use Jev.
ProjectbetaOptimises the instructions and criterion descriptions of a Jev Choice from labelled data with EvoPrompt or GEPA, scored by Brier.
ProjectbetaPython