Skip to content

alibaba/open-code-review

v1.0.0Apache-2.0

AI-powered code review on Git diffs — run ocr review as an agent skill: workspace changes, branch ranges, and single commits with structured, severity-rated findings.

Coding agent plugins

Open Code Review ships platform-specific integrations for Claude Code, Codex, Cursor, Kimi Code, and QCA Forward. Choose your platform below instead of adapting installation instructions written for a different agent.

All integrations require Git 2.41 or later. Install the ocr CLI first:

npm install -g @alibaba-group/open-code-review

Configure and test an OCR LLM before running a review, unless you plan to use Delegation Mode:

ocr config provider
ocr config model
ocr llm test

Claude Code

Run these commands inside Claude Code:

/plugin marketplace add alibaba/open-code-review
/plugin install open-code-review@open-code-review

This installs the /open-code-review:review and /open-code-review:delegate-review slash commands. See the Claude Code guide for manual installation, usage, and behavior.

Copilot CLI

Install the plugin, which adds the two Agent Skills open-code-review and open-code-review-delegate. Mention /open-code-review or just describe the task; Copilot picks the skill.

copilot plugin install alibaba/open-code-review:plugins/open-code-review

Copilot CLI's installer warns that direct installs (repo/path/URL) are deprecated in favor of plugin@marketplace; once this repository's marketplace index points at the plugin root, copilot plugin install <name>@open-code-review becomes the recommended form. Remove it with copilot plugin uninstall open-code-review. See the Copilot CLI guide for usage and manual installation.

Codex

Add this repository as a Codex marketplace, then start Codex:

codex plugin marketplace add alibaba/open-code-review
codex

Open /plugins, install and enable Open Code Review, then start a new task. The plugin exposes callable review skills backed by the local ocr CLI. For example:

@Open Code Review review my current changes
@Open Code Review review this branch against main
@Open Code Review review and fix high-confidence issues

The default review skill runs ocr review, which calls the LLM configured for ocr, not the Codex session. To review with your existing Codex quota and no OCR LLM setup, use the delegate skill. It runs ocr delegate preview and ocr delegate rule for file selection and rules, and Codex itself performs the review:

@Open Code Review delegate review my current changes

Cursor

This repository includes a Cursor plugin manifest at .cursor-plugin/plugin.json. For a local manual installation, copy the entire plugins/open-code-review/ directory to:

~/.cursor/plugins/local/open-code-review/

Verify that the manifest is located at ~/.cursor/plugins/local/open-code-review/.cursor-plugin/plugin.json, then restart Cursor or run Developer: Reload Window. The plugin provides the portable OCR review skills from the bundled skills/ directory.

See the Cursor plugin documentation for plugin loading and management details.

Kimi Code

This repository is itself a Kimi Code plugin. The manifest lives at the repository root (.kimi-plugin/plugin.json) because Kimi Code treats the repository root as the plugin root when installing from GitHub; its skills and commands paths point down into this directory.

Run this inside Kimi Code:

/plugins install https://github.com/alibaba/open-code-review

Then run /reload (or start a new session) to activate the plugin. It provides the /open-code-review:review and /open-code-review:delegate-review slash commands plus the portable OCR review skills, backed by the local ocr CLI.

Plugins are installed per-user and apply to every project. Manage the installation with /plugins list, /plugins info open-code-review, /plugins disable open-code-review, and /plugins remove open-code-review. See the Kimi Code plugin documentation for details.

QCA Forward

QCA Forward uses the host model through OCR delegation mode. Publish the open-code-review-delegate Skill, preinstall OCR in the referenced QCA environment, and create a Forward Template from the bundled example. No OCR LLM endpoint is required.

See the QCA Forward integration guide and template example.