What does Jev Prompt Enhancer do?

It is an agent skill that gives your coding agent a clearer task before it writes code: Jev scores your prompt and flags where the task, context, constraints or expected output are unclear, then the skill refines and rechecks before building.

How does the enhancement loop work?

Measure, Refine, Recheck, Build. Jev returns a quality score, the skill addresses the weak spots while preserving your intent, measures again, and proceeds when the prompt reaches the target. You can also measure a prompt directly with the measure.sh script.

When does the loop stop?

At 80/100 by default, or at a threshold you specify, or when all three enhancement attempts are used. At the limit the skill selects the highest-scoring prompt it produced.

How do I install and configure it?

Install with npx skills add joacotornello/prompt-enhancer. It needs Bash and Node.js 18 or newer (use Git Bash on Windows) with no npm dependencies, and JEV_API_KEY must be set in the environment used to launch your agent.

How is the API key read?

From the environment first; an existing nonempty environment key takes precedence. Otherwise the script reads literal single-line JEV_API_KEY assignments from .env, honoring quotes and trailing comments but never executing shell commands or variable expansion.

Can I run the tests without a Jev key?

Yes. npm test mocks Jev requests, so verification requires neither a real API key nor network access, according to the README.

Does it change what I asked for?

No. The README says your requirements and scope stay intact: the skill improves how you ask while keeping what you asked for at the center, and it applies to bug fixes, new features and review requests alike.

Jev directory

CommunityJev Agents & MCP

Jev Prompt Enhancer

An agent skill that measures your prompt with Jev before your coding agent starts work, refines the weak spots while preserving intent and scope, and rechecks until the prompt reaches a quality target — 80/100 by default, or three attempts.

See highlights Open github.com

Stars: 1
Language: Shell
Reading time: 2 min
Published by: Community
Author: joacotornello
Last verified: Oct 2026

joacotornello/jev-prompt-enhancer

User repository on GitHub

View on GitHub

Enhance your prompts using Jev by giving your coding agent a clearer task before it writes the first line.

Forks: 0
Last pushed: Updated Sep 2026

Repo stats from the GitHub API, cached Oct 2026.

Tagscommunity catalog-import logicrw agent-skill prompt-engineering

Works forJev for Scoring Jev for Decision gating

Used inJev for Software & DevTools

Languages: Shell
Platforms: CLI
Capabilities: Open sourceRequires API key

Highlights

Runs a Measure, Refine, Recheck, Build loop: Jev scores the prompt and pinpoints weak spots in the task, context, constraints or expected output.
Stops at 80/100 by default (or a threshold you set), or after three enhancement attempts; at the limit it keeps the highest-scoring prompt.
Installs with npx skills add joacotornello/prompt-enhancer; needs Bash and Node.js 18+ with no npm dependencies, and Git Bash on Windows.
Reads JEV_API_KEY from the environment or a .env file (JEV_ENV_FILE sets its path); only literal single-line assignments are parsed, never shell execution.
npm test verifies the skill with mocked Jev requests, so the checks need no real API key and no network access.

Quickstart

bash

npx skills add joacotornello/prompt-enhancer
export JEV_API_KEY='your-typesafe-api-key'

Watch out

Copy reviewed against the repo README on 2026-10-15; GitHub metadata lists no license for this repo. The 80/100 target and three-attempt cap are the README's documented defaults, and no benchmark of prompt-quality improvement is published.

Questions about this entry

Short answers drawn from the linked project or page.

From the Jev guides

Original write-ups on what this entry is for.

MCP tool routing: Jev in front of the tool call

Oct 16, 2026

Tool menus can outgrow the context that reads them. The catalog's Jev-based routers and gates narrow the list: benchmarked tool search, per-tool scoring, fail-closed approvals, and WebMCP in the browser.

Jev Agents & MCP#mcp#tool-routing#agents

Read article

Approval policies for coding agents: hooks, guards, and gateways