Skip to content
Metatext Decide — typed decisions for your agents, with a probability on every answerGet early access
Metatext
Sign Up

btx-skill-jev-judge

Claude Code plugin and Python CLI that hands judgments repeated over many items to Jev, with redaction, rate limits and a pilot run.

Agent skillbetaMIT
View on GitHub ↗

Overview

btx-skill-jev-judge is a Claude Code plugin with one skill, jev-judge, and the Python package behind it. When the same judgment repeats over many items, such as labeling 600 support tickets, Claude hands it to Jev through a tested command-line tool instead of reading every item into context or calling a large model per item.

The list summary highlights redaction, rate limiting, a pilot run and a check for flat questions (questions whose answers barely vary). Results are typed yes/no, choice or scale answers with probabilities, and each run reports its cost estimated from the per-million-token price in the code. The project has coverage, linting and security-scanning badges and is published on PyPI.

What the decision model decides

TypeQuestionAnswers
ChoiceWhich label fits this item (e.g. ticket category)?(inferred)—
Yes / noYes/no judgment repeated over each item(inferred)yes · no
ScorePosition on a scale for each item(inferred)—

How it works

  1. Claude recognizes a repeated judgment
  2. Run a pilot on a sample and check for flat questions
  3. Redact and send items to Jev with rate limiting
  4. Collect typed answers and a cost report

Strengths

  • Keeps items out of Claude's context
  • Pilot and flat-question checks catch bad questions early
  • Tested CLI with cost reporting

Caveats

  • Needs uv and a TypeSafe key
  • Cost estimate uses a hard-coded price

Getting started

/plugin marketplace add bitranox/btx-skill-jev-judge then /plugin install btx-skill-jev-judge@btx-skill-jev-judge

At a glance

Type
Agent skill
Status
beta
Language
Python
Runtime
Claude Code plugin + CLI
Has code
yes
Has tests
yes
License
MIT

Use cases and integrations

How curated lists describe it

  • Claude Code plugin and Python CLI: hand a judgment repeated over many items to…

    wh000wh000/awesome-jev-live

  • Claude Code plugin and Python CLI: hand a judgment repeated over many items to TypeSafe's Jev, with redaction, rate limiting, a pilot and a flat-question check.

    hellogumbo/awesome-jev

Frequently asked questions

When should Claude use it?

When the same judgment repeats over many items, like labeling hundreds of tickets.

What protects data?

Redaction before sending, per the listing.

Is cost reported?

Yes, each run reports an estimated cost.

Related projects

  • jevpipe

    Unix filter CLI that pipes files, lines or records to Jev for bulk yes/no, choice or score judgments so agents need not read everything.

    ProjectbetaPython
  • jev-organize

    Zero-dependency Node CLI that classifies a folder of company files with Jev by department, type, sensitivity, date, counterparty and PII.

    Agent skillbetaJavaScript
  • typesafe-jev-plugin

    Claude Code plugin that builds reusable Jev signatures for classification or scoring and runs CSV, Excel or text datasets through them.

    PluginbetaPython
  • Building with Jev★ 155

    Agent skill for writing and improving programs that call Jev: question design, state structure, thresholds and diagnosing wrong answers.

    Agent skillproduction
  • jev-mcp (freepik-company)

    Go MCP server giving any client typed Jev decisions via OpenRouter or TypeSafe, with validated answers and no paid retries.

    MCP serverproductionGo
  • askjev

    Unofficial hosted or local Jev MCP server giving agents a quick second opinion: it picks yes/no, scale or choice and returns probabilities.

    MCP serverbetaTypeScript