jev-code-reviewer README

The project's README from github.com/egma-ai/jev-code-reviewer. As one markdown file: README.md.

Jev-Code-Reviewer

Most PRs that are generated by agents today get YOLO merged because its hard for human mind to comprehend when your agent just suddenly shows up with 230 file changes. This is an attempt to reduce the mental burden by classifying each change in a review to P0, P1, P2. Only P0 are shown by default. The priorities are configurable. The diffs are also show using a natural language. The original code is one toggle away.

It runs on your computer, for reviewing your own coding agent's PRs, and posts nothing to GitHub.

Jev-Code-Reviewer on the demonstration PR

Watch the recording · Open the demo PR. The bundled recording uses real Jev classifications and labeled, prepared explanation copy (provenance).

Quick demo

Needs Node.js 22+.

git clone https://github.com/egma-ai/jev-code-reviewer.git
cd jev-code-reviewer
npm install
npm run demo

Open http://127.0.0.1:4731/demo for the replay, which makes no provider calls, or load the extension (below) and open the demo PR's Files changed page.

Review your own PR

Needs the GitHub CLI signed in (gh auth login) and a local clone whose origin is the PR's repository.

npm link                 # adds the jev-reviewer command
jev-reviewer setup       # stores your TypeSafe and OpenAI keys; run it in your own terminal
jev-reviewer doctor      # checks tools, key sources, the local server, and provider access
jev-reviewer serve       # leave running
jev-reviewer analyze --pr https://github.com/OWNER/REPO/pull/123 --repo /path/to/clone

Then in Chrome:

  1. Open chrome://extensions, turn on Developer mode, click Load unpacked, and pick this repo's extension/ folder.
  2. Run jev-reviewer token | pbcopy. In the extension popup, open Connection, paste, and click Save. It should say Paired.
  3. Open the PR's Files changed page (/pull/<n>/files) and turn on Show logic in place of code.

Optional: uv tool install graphifyy adds a local code graph for better context. To have your coding agent run analyze after it opens a PR, install the agent skill.

Good to know

Development

npm test
node scripts/test-extension.mjs   # needs Chromium and port 4731 free

See architecture, demo guide, and extension details.

MIT.

jev-code-reviewer (egma-ai/jev-code-reviewer)

Searching Database…

144 ·GitHub
Jev-code-reviewer
这是什么?Jev 是一款面向编程智能体所开拉取请求的本地代码评审工具:它把每个 diff 块按 P0、P1 或 P2 的关注度优先级分类,并用自然语言而非原始代码来解释变更。
这是命令行工具、浏览器扩展,还是两者都是?它由三部分协同组成:分析 diff 的本地 Node.js 命令行工具、把解释叠加到 GitHub Files changed 页面上的 Chrome 扩展,以及你的编程智能体在开完 PR 后可以调用的智能体技能。
怎么用它评审我自己的 PR?把 extension/ 文件夹作为未打包的 Chrome 扩展加载,运行 'jev-reviewer token | pbcopy' 并粘贴到扩展的 Connection 弹窗里,然后打开 PR 的 Files changed 页面,开启 'Show logic in place of code'。
试用之前需要装好什么?需要 Node.js 22+ 和已登录的 GitHub CLI(gh auth login),以及一个 origin 与 PR 仓库一致的本地克隆;分析还需要 OpenAI API 密钥。
它会分析整个 diff 吗,在任何 GitHub PR 页面都能用吗?不能:默认只分析前 12 个 diff 块(用 --max-units 最多可到 100 个),未分析的文件保持 GitHub 原始代码视图,而且它只在经典 Files changed 页面上工作,不支持 GitHub 新的 /changes 页面。
我的代码会被发送到别处吗?会,分析步骤会把变更代码及上下文发送给 TypeSafe 和 OpenAI;Chrome 扩展本身只与 127.0.0.1:4731 上的本地服务器通信。
P0、P1、P2 优先级是怎么定的,我能改吗?它们表示一个变更值得多少人工关注,而不是正确性。P0 文件默认展开,P1 和 P2 默认折叠;你可以把 config/policy.json 复制成被评审仓库里的 .jev-reviewer.json 文件,或传入 --policy 来重新调整阈值。
我的 API 密钥存在哪里?密钥存放在 ~/.config/jev-reviewer/credentials.json(仅所有者可读、未加密),而不是 shell 导出的环境变量,因为编程智能体的 shell 常常看不到环境变量。可以用 'node scripts/setup-keys.mjs --replace-openai' 更换 OpenAI 密钥。
不接入自己的 PR 能试用吗?可以,内置了一段演示录像,使用真实的 Jev 分类和准备好的解释文案,回放时不会调用任何服务商;另外还有一个开放的演示 PR,可以加载扩展来体验。
我运行 analyze 之后 PR 又有新提交,会怎样?旧报告绝不会自动显示;扩展会一直显示 GitHub 的原始代码,直到你重新运行 analyze 并点击 Refresh report。
What are similar projects?perch, pr-agent, no-mistakes, 1code, open-swe, open-code-review, issue-graph, diffle
Which topics is it in?pull-request, byok, devtools, code-quality, code-review, chrome-extension, browser-extension, github, cli-tool, git, local-first, coding-agent
Creator
egma-ai
Size
4 MB
Stars
144
Updated
CommitOct 05
Age
29d
Commits · 90 Days