Will it read my repository on its own, and what context gets sent?

No. Jev Review never reads the repository automatically; the caller passes task, diff, files, repositoryContext and an optional previousEvaluation, so the agent decides what to submit. It adds no character, token or file-count limit of its own, and it has no hosted backend, database or telemetry service.

Why is there no single overall quality score out of 100?

Deliberate. The README argues that dimension changes such as Readability 6.3 to 8.1 and Security 8.2 to 8.2 are more useful than a blended percentage, so there is no synthetic 82/100. You get an independent 1 to 10 score and 0 to 1 confidence per dimension, plus per-metric deltas when you pass a previous evaluation.

Will it explain why a score is low, or do I still have to do the review?

It will not. Jev returns typed Score, Choice and Noul decisions rather than a free-form review essay and does not generate a root-cause explanation. The server converts those into prioritized weak dimensions and coarse predefined rubric hints; your coding agent remains responsible for diagnosing the weakness and changing the code.

Will a large diff get rejected for being too big?

Not by this server, but the Jev API enforces its own ceiling on the submitted state. Live jev-latest behavior indicates roughly 32,768 tokens, though that figure is not in the published API documentation or OpenAPI schema and may change; on max_tokens_exceeded the server asks the agent to trim unrelated context or split the change into coherent review slices.

What fields does the jev_review tool accept?

task, diff, files (each with path and content), repositoryContext, and previousEvaluation. At least one current-context field is required, and callers normally send the task and a focused diff.

What is the token ceiling for a review?

The Jev API currently accepts roughly 32,768 tokens of submitted state, though the number is unpublished and may change. On max_tokens_exceeded the server asks you to trim unrelated context or split the change into coherent review slices.

How many quality dimensions does Jev Review score?

Fifteen are always evaluated when the context is sufficient, from correctness and cognitive complexity to security and documentation. Four more, such as performance, scalability, compatibility, and observability, are scored only when relevant evidence is present.

Jev directory

CommunityJev Tools & CLIs

Jev Review — NiazMorshed2007

A local-first MCP server that gives Claude Code, Codex, Cursor, and OpenCode structured quality scores from Jev across correctness, complexity, tests, security, and other dimensions while the agent writes code.

See highlights Open github.com

Stars: 243
Language: TypeScript
License: MIT
Reading time: 3 min
Published by: Community
Author: NiazMorshed2007
Last verified: Sep 2026

NiazMorshed2007/jev-review

User repository on GitHub

View on GitHub

Local-first MCP plugin for continuous software-quality review by AI coding agents, powered by Jev.

Forks: 25
Last pushed: Updated Sep 2026

Repo stats from the GitHub API, cached Oct 2026.

This is one of several Jev listings for the same project. The canonical entry is Jev Review — devagrawal09.

About this project

Continuous software-quality review for AI coding agents, powered by Jev. Quick start · Client setup · Quality dimensions · Security. Jev Review runs as a local MCP server and gives Claude Code, Codex, Cursor, and OpenCode structured quality scores while they work. Your coding agent remains responsible for diagnosing weaknesses and changing the code; Jev supplies a fast scalar signal across correctness, complexity, changeability, modularity, tests, security, and other independent quality dimensions.

Quoted from the project's README. Read the full README.

Tagscommunity typescript mcp coding-agent claude-code codex

Works forJev for Code Jev for Tool calling Jev for Scoring

Used inJev for Security & Privacy Jev for Data & Analytics

Highlights

One focused MCP tool, jev_review, runs as a local Node.js process over stdio; there is no hosted backend, database, or telemetry service.
Jev returns typed Score, Choice, and Noul decisions, which the server converts into dimension metrics, confidence levels, and previous-evaluation comparisons.
There is deliberately no blended 82/100 score; changes are reported per dimension, such as Readability 6.3 to 8.1.
The coding agent stays responsible for diagnosing weaknesses and editing code; Jev only supplies the scalar signal.
Install from GitHub with npx plugins add; manual MCP configuration is documented for all four clients.

Quickstart

bash

export JEV_API_KEY="YOUR_API_KEY"
npx plugins add NiazMorshed2007/jev-review

Watch out

MIT-licensed. Needs Node.js 20+, a Jev API key, and one of the four supported clients; review context (task, diff, files, repositoryContext) is sent to TypeSafe's API, which enforces roughly a 32,768-token state ceiling.

Questions about this entry

Short answers drawn from the linked project or page.

Reactions & coverage

Posts, threads, and videos about this entry from around the web.

X: jev-review: a local-first score loop for coding agents

Reddit: One command to wire Jev into Claude Code, Claude Desktop, and Codex.

YouTube: Ray Amjad puts Jev inside an agentic coding loop and looks at what it costs to run.

YouTube: RoboNuggets puts Jev to work inside Claude Code.

From the Jev guides

Original write-ups on what this entry is for.

Putting Jev in the Coding Loop: MCP Servers, Gates, and Skills

Sep 14, 2026

The fastest-growing corner of the Jev ecosystem is coding agents. A tour of the MCP servers, tool-call gates, context pruners, and skill rankers built for that loop.

Jev Tools & CLIs#agents#mcp#coding-agent

Read article

Jev at home: voice pipelines, Assist routing, and sentinels

Oct 16, 2026

How Home Assistant conversation agents, macOS voice pipelines, and an advisory sentinel put a fast typed decision in the loop — one call per utterance, thresholds in code, and a daily token budget.

Jev Tools & CLIs#home-assistant#voice#automation

Read article

Search visibility, audited by decisions

Oct 16, 2026

Jev-based SEO tools split the crawl from the call: deterministic crawlers count and parse, while typed judgments rank fixes, place internal links, score search results, and track AI-answer mentions — with costs and caveats attached.

Jev Tools & CLIs#seo#search-visibility#audits

Read article

From the Jev AI model community

Posts from builders shipping with Jev right now.

Follow @typesafeai

Jev Review — NiazMorshed2007

About this project

Highlights

Quickstart

Questions about this entry

Will it read my repository on its own, and what context gets sent?

Why is there no single overall quality score out of 100?

Will it explain why a score is low, or do I still have to do the review?

Will a large diff get rejected for being too big?

What fields does the jev_review tool accept?

What is the token ceiling for a review?

How many quality dimensions does Jev Review score?

Reactions & coverage

Related terms

More Jev resources like this

From the Jev guides

From the Jev AI model community