jev-skill (marcodicesare-dev)
Community Jev skill for Claude Code and Codex: a guide, a dependency-free Python CLI and 5 measured question sets.
Overview
This jev-skill by marcodicesare-dev is an independent community skill for Claude Code, Codex and other coding agents. It grew out of the author's attempt to make an AI writer cheaper with Jev, which made the writing worse, followed by testing to find where the model does help. The research tally records 19,367 successful Jev API responses and 90.4 million metered input tokens over 4 days, including retries and reruns.
The repository contains a long guide (also published as a web page) with a decision map and a first workflow, a standard-library Python CLI, 5 measured question sets, and notes on failures. Its suggested first prompt asks the agent to find one repeated, bounded decision in a workflow and show the question, state and a small labelled test before changing anything. The quick start routes a customer message to billing, account or other, and the README notes that routing does not verify a charge or approve a refund. The CLI can call TypeSafe directly or through other gateways. It points readers to TypeSafe's official skill for current provider behavior.
What the decision model decides
| Type | Question | Answers |
|---|---|---|
| Choice | Which team should see this customer message first? | billing · account · other |
How it works
- Clone into your agent's skills folder
- Set TYPESAFE_API_KEY locally
- Ask the agent to find one bounded decision
- Build a small labelled test
- Only then change the workflow
Strengths
- Grounded in a large personal test log
- Includes failures, not just wins
- No package dependencies
Caveats
- Independent, not the official skill
- Findings come from one author's workloads
- Some routes need other gateway keys
Getting started
git clone https://github.com/marcodicesare-dev/jev-skill.git ~/.claude/skills/jev and export TYPESAFE_API_KEYAt a glance
- Type
- Agent skill
- Status
- beta
- Language
- Python
- Runtime
- Agent skill and CLI
- Has code
- yes
- License
- MIT
Use cases and integrations
How curated lists describe it
Jev skill for Claude Code and Codex: TypeSafe System One guide, Python CLI, 5 tested recipes, and findings from 19,367 Jev calls.
logicrw/awesome-jev-projectsJev skill for Claude Code and Codex: TypeSafe System One guide, Python CLI, 5 tested recipes, and findings from 19,367 Jev calls.
hellogumbo/awesome-jev
Frequently asked questions
What does the skill help with?
Finding repeated, bounded decisions that suit Jev and testing them before changing a workflow.
How much testing is behind it?
19,367 successful Jev responses and 90.4 million input tokens over 4 days, per its tally.
Is it TypeSafe's official skill?
No. It is an independent community project and links to the official one.
Related projects
- 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 Portable Agent Skill teaching coding agents to use Jev as a bounded judgment layer, plus an optional MCP bridge.
MCP serverbetaTypeScriptIndependent agent skill that makes a coding agent design, build, test and tune a bounded Jev decision layer with a contract first.
Agent skillbetaClaude Code and Codex skills plus a study course for building with Jev, including Jevaluate, which rates how well projects use Jev.
ProjectbetaAgent skill and CLI that chains Jev question design: find decision calls, check fit, draft, lint, calibrate thresholds and plan a cascade.
Agent skillprototypePython- typesafe-ai/skills★ 2,685
Official TypeSafe agent skill and Claude Code plugin that teaches coding agents to design System One workflows with typed judgments.
Agent skillproduction