### Install Dependencies and Start Watch Mode

Source: https://github.com/alibaba/open-code-review/blob/main/extensions/vscode/README.md

Navigate to the VS Code extension directory, install dependencies using Yarn, and start the watch-mode development build.

```bash
cd extensions/vscode
yarn install
yarn watch
```

--------------------------------

### Configure LLM Interactively

Source: https://github.com/alibaba/open-code-review/blob/main/README.md

Guides the user through selecting an LLM provider and model. This is the recommended method for initial setup.

```bash
ocr config provider          # Select a built-in provider or add a custom one
ocr config model             # Pick a model for the active provider
```

--------------------------------

### Install CLI from Source

Source: https://github.com/alibaba/open-code-review/blob/main/README.md

Clones the repository, builds the project, and installs the binary to /usr/local/bin/ocr.

```bash
git clone https://github.com/alibaba/open-code-review.git
cd open-code-review
make build
sudo cp dist/opencodereview /usr/local/bin/ocr
```

--------------------------------

### Install CLI from GitHub Release (macOS/Linux)

Source: https://github.com/alibaba/open-code-review/blob/main/README.md

Installs the latest binary for your OS/architecture using a curl script. The script verifies the checksum and installs 'ocr' in /usr/local/bin.

```bash
curl -fsSL https://raw.githubusercontent.com/alibaba/open-code-review/main/install.sh | sh
```

--------------------------------

### Setup Workflow File

Source: https://github.com/alibaba/open-code-review/blob/main/examples/github_actions/README.md

Copy the provided workflow file to your repository's .github/workflows/ directory to enable OpenCodeReview integration.

```bash
mkdir -p .github/workflows
cp ocr-review.yml .github/workflows/ocr-review.yml
```

--------------------------------

### Install CLI from GitHub Release with Custom Directory/Version

Source: https://github.com/alibaba/open-code-review/blob/main/README.md

Installs the CLI from a GitHub release, allowing customization of the installation directory and specific version.

```bash
OCR_INSTALL_DIR="$HOME/.local/bin" OCR_VERSION=v1.3.13 \
  sh -c "$(curl -fsSL https://raw.githubusercontent.com/alibaba/open-code-review/main/install.sh)"
```

--------------------------------

### Build and Test Project

Source: https://github.com/alibaba/open-code-review/blob/main/CONTRIBUTING.md

Follow these commands to build the project and run tests after cloning the repository. Ensure you have Go, Git, and Make installed.

```bash
git clone https://github.com/<your-username>/open-code-review.git
cd open-code-review
git remote add upstream https://github.com/alibaba/open-code-review.git
make build
make test
```

--------------------------------

### Install OCR as a Codex Plugin (Marketplace)

Source: https://github.com/alibaba/open-code-review/blob/main/README.md

Commands to install the OCR plugin for local Codex from the marketplace.

```bash
codex plugin marketplace add alibaba/open-code-review
codex
/plugins
```

--------------------------------

### Commit Message Examples

Source: https://github.com/alibaba/open-code-review/blob/main/CONTRIBUTING.md

Examples demonstrating the Conventional Commits format for various types of changes.

```text
feat(agent): add support for custom tool definitions
```

```text
fix(llm): handle timeout errors in Anthropic API calls
```

```text
docs(README): update configuration examples
```

--------------------------------

### Install OCR CLI

Source: https://github.com/alibaba/open-code-review/blob/main/skills/open-code-review/SKILL.md

Installs the open-code-review CLI globally using npm. Run this if the 'ocr' command is not found.

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

--------------------------------

### Manual Download and Install CLI (Linux x86_64)

Source: https://github.com/alibaba/open-code-review/blob/main/README.md

Manually downloads the binary for Linux on x86_64 architecture, makes it executable, and moves it to /usr/local/bin.

```bash
# Linux (x86_64)
curl -Lo ocr https://github.com/alibaba/open-code-review/releases/latest/download/opencodereview-linux-amd64
chmod +x ocr && sudo mv ocr /usr/local/bin/ocr
```

--------------------------------

### Install Extension Locally

Source: https://github.com/alibaba/open-code-review/blob/main/extensions/vscode/README.md

Use this command to install the generated .vsix file locally. Alternatively, you can install it via the VS Code Extensions panel by selecting 'Install from VSIX…'.

```bash
code --install-extension open-code-review-vscode-<version>.vsix
```

--------------------------------

### Manual Download and Install CLI (Linux ARM64)

Source: https://github.com/alibaba/open-code-review/blob/main/README.md

Manually downloads the binary for Linux on ARM64 architecture, makes it executable, and moves it to /usr/local/bin.

```bash
# Linux (ARM64)
curl -Lo ocr https://github.com/alibaba/open-code-review/releases/latest/download/opencodereview-linux-arm64
chmod +x ocr && sudo mv ocr /usr/local/bin/ocr
```

--------------------------------

### Install OCR CLI Globally

Source: https://github.com/alibaba/open-code-review/blob/main/extensions/vscode/README.md

Install the Open Code Review CLI tool globally using npm.

```bash
npm i -g @alibaba-group/open-code-review
```

--------------------------------

### Manual Download and Install CLI (macOS Intel)

Source: https://github.com/alibaba/open-code-review/blob/main/README.md

Manually downloads the binary for macOS on Intel processors, makes it executable, and moves it to /usr/local/bin.

```bash
# macOS (Intel)
curl -Lo ocr https://github.com/alibaba/open-code-review/releases/latest/download/opencodereview-darwin-amd64
chmod +x ocr && sudo mv ocr /usr/local/bin/ocr
```

--------------------------------

### Manual Download and Install CLI (Windows x86_64)

Source: https://github.com/alibaba/open-code-review/blob/main/README.md

Manually downloads the Windows x86_64 binary. The user needs to move 'ocr.exe' to a directory in their PATH.

```bash
# Windows (x86_64) — move ocr.exe to a directory in your PATH
curl -Lo ocr.exe https://github.com/alibaba/open-code-review/releases/latest/download/opencodereview-windows-amd64.exe
```

--------------------------------

### Manual Download and Install CLI (Windows ARM64)

Source: https://github.com/alibaba/open-code-review/blob/main/README.md

Manually downloads the Windows ARM64 binary. The user needs to move 'ocr.exe' to a directory in their PATH.

```bash
# Windows (ARM64) — move ocr.exe to a directory in your PATH
curl -Lo ocr.exe https://github.com/alibaba/open-code-review/releases/latest/download/opencodereview-windows-arm64.exe
```

--------------------------------

### Install OCR as a Codex Plugin (Local Checkout)

Source: https://github.com/alibaba/open-code-review/blob/main/README.md

Commands to install the OCR plugin for local Codex from a local checkout or fork.

```bash
codex plugin marketplace add .
codex
/plugins
```

--------------------------------

### Install OCR as a Skill using npx

Source: https://github.com/alibaba/open-code-review/blob/main/README.md

Use this command to install the OCR skill into your project, enabling your coding agent to invoke OCR for code reviews.

```bash
npx skills add alibaba/open-code-review --skill open-code-review
```

--------------------------------

### Manual Download and Install CLI (macOS Apple Silicon)

Source: https://github.com/alibaba/open-code-review/blob/main/README.md

Manually downloads the binary for macOS on Apple Silicon, makes it executable, and moves it to /usr/local/bin.

```bash
# macOS (Apple Silicon)
curl -Lo ocr https://github.com/alibaba/open-code-review/releases/latest/download/opencodereview-darwin-arm64
chmod +x ocr && sudo mv ocr /usr/local/bin/ocr
```

--------------------------------

### GitLab CI Workflow Example

Source: https://github.com/alibaba/open-code-review/blob/main/examples/gitlab_ci/README.md

Illustrates the automated workflow triggered by Merge Request events, leading to code review and discussion posting.

```text
MR Created/Updated → GitLab Pipeline Triggered → OCR Reviews Diff → Discussions Posted on MR
```

--------------------------------

### Invoke OCR in Codex

Source: https://github.com/alibaba/open-code-review/blob/main/README.md

Examples of how to invoke the installed Open Code Review plugin in Codex for various review scenarios.

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

--------------------------------

### Check OCR CLI Installation and LLM Connectivity

Source: https://github.com/alibaba/open-code-review/blob/main/skills/open-code-review/SKILL.md

Verifies if the 'ocr' CLI is installed and if it can connect to the configured LLM. Run these commands before initiating a code review.

```bash
which ocr || echo "NOT INSTALLED"

ocr llm test
```

--------------------------------

### Rule File Format Example

Source: https://github.com/alibaba/open-code-review/blob/main/README.md

Defines rules for code review, specifying file paths and associated rule descriptions. Supports glob patterns like ** and brace expansion.

```json
{
  "rules": [
    {
      "path": "force-api/**/*.java",
      "rule": "All new methods must validate required parameters for null values"
    },
    {
      "path": "**/*mapper*.xml",
      "rule": "Check SQL for injection risks, parameter errors, and missing closing tags"
    }
  ]
}
```

--------------------------------

### Install OCR as a Claude Code Plugin

Source: https://github.com/alibaba/open-code-review/blob/main/README.md

Commands to install the OCR plugin for Claude Code, registering the \/open-code-review:review slash command.

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

--------------------------------

### Configure GitLab CI for OpenCodeReview

Source: https://github.com/alibaba/open-code-review/blob/main/examples/gitlab_ci/README.md

This snippet shows how to install OpenCodeReview, configure its settings, and use a Python wrapper script to conditionally run reviews based on existing comments. It helps prevent re-reviewing MRs on every push.

```yaml
script:
  # Install OpenCodeReview
  - npm install -g @alibaba-group/open-code-review

  # Configure OCR
  - mkdir -p ~/.open-code-review
  - |
    ocr config set llm.url $OCR_LLM_URL
    ocr config set llm.auth_token $OCR_LLM_AUTH_TOKEN
    ocr config set llm.model $OCR_LLM_MODEL
    ocr config set llm.use_anthropic false
    ocr config set llm.extra_body '{"thinking": {"type": "disabled"}}'

  # Check for existing OCR reviews and run review only if not found
  - |
    python3 << 'WRAPPER_SCRIPT'
    import json
    import os
    import subprocess
    import sys
    import urllib.request

    GITLAB_URL = os.environ.get("CI_SERVER_URL", "https://gitlab.com")
    PROJECT_ID = os.environ["CI_PROJECT_ID"]
    MR_IID = os.environ["CI_MERGE_REQUEST_IID"]
    API_TOKEN = os.environ["GITLAB_API_TOKEN"]
    SOURCE_BRANCH = os.environ["CI_MERGE_REQUEST_SOURCE_BRANCH_NAME"]
    TARGET_BRANCH = os.environ["CI_MERGE_REQUEST_TARGET_BRANCH_NAME"]

    # Check for existing OCR reviews
    url = f"{GITLAB_URL}/api/v4/projects/{PROJECT_ID}/merge_requests/{MR_IID}/notes?per_page=100"
    req = urllib.request.Request(url, headers={"PRIVATE-TOKEN": API_TOKEN})
    with urllib.request.urlopen(req) as resp:
        notes = json.loads(resp.read().decode("utf-8"))

    for note in notes:
        if "OpenCodeReview" in note.get("body", ""):
            print("⏭️ OCR has already reviewed this MR. Skipping to save tokens.")
            print("Delete previous OCR comments to re-trigger review.")
            sys.exit(0)

    # No existing review found - run OCR
    print("🔍 No existing OCR review found. Running review...")
    COMMIT_SHA = os.environ["CI_COMMIT_SHA"]
    result = subprocess.run([
        "ocr", "review",
        "--from", f"origin/{TARGET_BRANCH}",
        "--to", COMMIT_SHA,
        "--format", "json",
        "--audience", "agent"
    ], capture_output=True, text=True)

    # Save output for the posting script
    with open("/tmp/ocr-result.json", "w") as f:
        f.write(result.stdout)
    with open("/tmp/ocr-stderr.log", "w") as f:
        f.write(result.stderr)

    print("OCR review completed.")
    WRAPPER_SCRIPT

  # Post review comments to MR
  - |
    python3 << 'PYTHON_SCRIPT'
    ...existing post script...
    PYTHON_SCRIPT
```

--------------------------------

### GitHub Suggestion Syntax Example

Source: https://github.com/alibaba/open-code-review/blob/main/examples/github_actions/README.md

Demonstrates how OCR uses GitHub's native suggestion syntax for inline code review comments. Reviewers can apply suggested fixes with a single click.

```markdown
**Suggestion:**
```suggestion
// Fixed code here
```
```

--------------------------------

### Install Specific OCR Version

Source: https://github.com/alibaba/open-code-review/blob/main/examples/gitlab_ci/README.md

Install a specific version of the Open Code Review package using npm.

```yaml
script:
  - npm install -g @alibaba-group/open-code-review@1.0.0
```

--------------------------------

### Install Specific Open Code Review Version

Source: https://github.com/alibaba/open-code-review/blob/main/examples/github_actions/README.md

Specify the exact version of the OpenCodeReview package to install using npm. This ensures consistent behavior across runs.

```yaml
- name: Install OpenCodeReview
  run: npm install -g @alibaba-group/open-code-review@1.0.0
```

--------------------------------

### Run Code Review with Business Context

Source: https://github.com/alibaba/open-code-review/blob/main/plugins/open-code-review/skills/open-code-review/SKILL.md

Execute the OCR review command, providing business context via the --background flag for improved review quality. Always use --audience agent to suppress UI and get only the final summary.

```bash
ocr review --audience agent --background "business context here" [user-args]
```

--------------------------------

### Use Custom Review Rules with OCR

Source: https://github.com/alibaba/open-code-review/blob/main/README.md

Apply custom review rules defined in a JSON file to guide the code review process.

```bash
ocr review --rule /path/to/my-rules.json
```

--------------------------------

### Build Production Artifacts

Source: https://github.com/alibaba/open-code-review/blob/main/extensions/vscode/README.md

Run this command to create production-ready artifacts for the extension host and webview. The output files are located in the 'out/' directory.

```bash
yarn build        # production build (webpack production)
```

--------------------------------

### Configure OCR Provider and Model

Source: https://github.com/alibaba/open-code-review/blob/main/README.md

Use these commands to interactively set up the code review provider and model. You can also list available LLM providers.

```bash
ocr config provider
ocr config model
ocr llm providers
```

--------------------------------

### Preview Files to be Reviewed

Source: https://github.com/alibaba/open-code-review/blob/main/plugins/open-code-review/skills/open-code-review/SKILL.md

Use the --preview or -p flag to see which files would be included in the review process without actually running the LLM analysis. This is useful for dry-runs.

```bash
ocr review --preview
```

--------------------------------

### Configure LLM Settings via CLI

Source: https://github.com/alibaba/open-code-review/blob/main/extensions/vscode/README.md

Configure Language Model settings such as URL, authentication token, model name, and provider using the OCR CLI.

```bash
ocr config set llm.url https://api.anthropic.com/v1/messages
```

```bash
ocr config set llm.auth_token sk-...
```

```bash
ocr config set llm.model claude-opus-4-6
```

```bash
ocr config set llm.use_anthropic true
```

--------------------------------

### Run Code Review with Business Context

Source: https://github.com/alibaba/open-code-review/blob/main/skills/open-code-review/SKILL.md

Initiates a code review using the 'ocr review' command. It's recommended to provide business context via the --background flag for improved review quality. The --audience agent flag suppresses UI and outputs only the summary.

```bash
ocr review --audience agent --background "business context here" [user-args]
```

--------------------------------

### Copying GitLab CI Pipeline File

Source: https://github.com/alibaba/open-code-review/blob/main/examples/gitlab_ci/README.md

Demonstrates how to copy the OpenCodeReview GitLab CI pipeline configuration file to your project's repository.

```bash
cp .gitlab-ci.yml /path/to/your/repo/.gitlab-ci.yml
```

--------------------------------

### Run Development Scripts

Source: https://github.com/alibaba/open-code-review/blob/main/extensions/vscode/README.md

Available scripts for development builds, testing, linting, and packaging the VS Code extension.

```bash
yarn compile
yarn watch
yarn build
yarn test
yarn lint
yarn package
```

--------------------------------

### Configure Telemetry Settings

Source: https://github.com/alibaba/open-code-review/blob/main/README.md

Use these commands to enable and configure OpenTelemetry integration for observability. Set the exporter to 'otlp' and specify the endpoint.

```bash
ocr config set telemetry.enabled true
ocr config set telemetry.exporter otlp
ocr config set telemetry.otlp_endpoint localhost:4317
```

--------------------------------

### Perform OCR Review with Default Settings

Source: https://github.com/alibaba/open-code-review/blob/main/README.md

Initiate a code review of your workspace changes using the default configuration.

```bash
ocr review
```

--------------------------------

### Perform Code Review (Workspace Mode)

Source: https://github.com/alibaba/open-code-review/blob/main/README.md

Reviews all staged, unstaged, and untracked changes in the current project directory.

```bash
cd your-project

# Workspace mode — review all staged, unstaged, and untracked changes
ocr review
```

--------------------------------

### Including GitLab CI Pipeline File

Source: https://github.com/alibaba/open-code-review/blob/main/examples/gitlab_ci/README.md

Shows how to include the OpenCodeReview GitLab CI pipeline configuration using GitLab's `include` feature in an existing `.gitlab-ci.yml` file.

```yaml
include:
  - local: 'ci_demo/gitlab_ci/.gitlab-ci.yml'
```

--------------------------------

### Preview Files to be Reviewed

Source: https://github.com/alibaba/open-code-review/blob/main/skills/open-code-review/SKILL.md

Performs a dry-run to preview which files would be reviewed without invoking the LLM. Use the --preview flag.

```bash
ocr review --preview
```

--------------------------------

### Package Release (.vsix)

Source: https://github.com/alibaba/open-code-review/blob/main/extensions/vscode/README.md

This command builds a release package (.vsix) for the extension. It first triggers a production build, then excludes unnecessary files, and finally produces the .vsix file in the current directory. The packaging tool used is @vscode/vsce.

```bash
yarn package      # = vsce package --no-yarn
```

--------------------------------

### Code Quality Checks

Source: https://github.com/alibaba/open-code-review/blob/main/CONTRIBUTING.md

Run these commands to ensure code quality, format the code, lint, and test the project. These checks should pass before submitting changes.

```bash
# Format and lint (Go standard tooling)
go fmt ./...
go vet ./...

# Run tests with race detection
make test

# Build successfully
make build
```

--------------------------------

### Preview Rule Application

Source: https://github.com/alibaba/open-code-review/blob/main/plugins/open-code-review/skills/open-code-review/SKILL.md

Checks which rule applies to a specific file before initiating a review. Ensure the file path is correct.

```bash
ocr rules check src/main/java/com/example/Foo.java
```

--------------------------------

### Preview Files for OCR Review

Source: https://github.com/alibaba/open-code-review/blob/main/README.md

Before performing a full review, preview which files will be analyzed without making any LLM calls. This is useful for understanding the scope of the review.

```bash
ocr review --preview
ocr review -c abc123 -p
```

--------------------------------

### Secure OCR Viewer with Host Header Allowlist

Source: https://github.com/alibaba/open-code-review/blob/main/README.md

Configure the OCR viewer to serve session data securely by enforcing a Host-header allowlist. This prevents DNS-rebinding attacks. Non-loopback hostnames must be added via the OCR_VIEWER_ALLOWED_HOSTS environment variable.

```bash
OCR_VIEWER_ALLOWED_HOSTS=review.internal,ocr.lan ocr viewer --addr :3000
```

--------------------------------

### Review Branch Comparison

Source: https://github.com/alibaba/open-code-review/blob/main/skills/open-code-review/SKILL.md

Compares two branches (e.g., 'main' and a feature branch) for code review. Provide business context with -b and use --audience agent for agent-only output.

```bash
ocr review --audience agent -b "context" --from main --to <branch>
```

--------------------------------

### Test LLM Connectivity

Source: https://github.com/alibaba/open-code-review/blob/main/README.md

Runs a test to verify that the configured LLM is reachable and responsive.

```bash
ocr llm test
```

--------------------------------

### Configure LLM Auth Header (Anthropic)

Source: https://github.com/alibaba/open-code-review/blob/main/README.md

Sets the HTTP header for the API key when using Anthropic. Defaults to 'authorization', but should be 'x-api-key' for standard keys.

```bash
ocr config set llm.auth_header x-api-key
```

--------------------------------

### Review Working Copy Changes

Source: https://github.com/alibaba/open-code-review/blob/main/skills/open-code-review/SKILL.md

Reviews staged, unstaged, and untracked changes in the working directory. Use the -b flag for business context and --audience agent for agent-only output.

```bash
ocr review --audience agent -b "context"
```

--------------------------------

### Configure LLM using Environment Variables

Source: https://github.com/alibaba/open-code-review/blob/main/skills/open-code-review/SKILL.md

Sets environment variables for LLM configuration, including URL, token, model, and Anthropic usage. This method has the highest priority and is recommended for CI environments.

```bash
export OCR_LLM_URL=https://api.anthropic.com/v1/messages
export OCR_LLM_TOKEN=<api-key>
export OCR_LLM_MODEL=claude-opus-4-6
export OCR_USE_ANTHROPIC=true
```

--------------------------------

### Configure LLM via Environment Variables

Source: https://github.com/alibaba/open-code-review/blob/main/README.md

Sets LLM configuration using environment variables, which take the highest priority. Also compatible with Claude Code environment variables.

```bash
export OCR_LLM_URL=https://api.anthropic.com/v1/messages
export OCR_LLM_TOKEN=your-api-key-here
export OCR_LLM_MODEL=claude-opus-4-6
export OCR_USE_ANTHROPIC=true
```

--------------------------------

### Custom Rule File Format

Source: https://github.com/alibaba/open-code-review/blob/main/plugins/open-code-review/skills/open-code-review/SKILL.md

Defines project-specific rules for code review. Use glob patterns for file paths and descriptive strings for rules.

```json
{
  "rules": [
    {
      "path": "**/*.java",
      "rule": "All new methods must validate required parameters for null"
    },
    {
      "path": "**/*mapper*.xml",
      "rule": "Check SQL for injection risks and missing closing tags"
    }
  ]
}
```

--------------------------------

### Configure Retry and Delay Settings

Source: https://github.com/alibaba/open-code-review/blob/main/examples/github_actions/README.md

Adjust repository variables like `OCR_RETRY_BASE_DELAY`, `OCR_RETRY_MAX_DELAY`, and `OCR_MAX_RETRIES` to fine-tune the rate-limit handling and retry behavior when posting review comments individually.

```yaml
# Example of setting variables in GitHub Actions workflow (not directly in code snippet, but relevant context)
# These are configured in repository settings -> Actions -> Variables
```

--------------------------------

### Provide Requirement Context for OCR Review

Source: https://github.com/alibaba/open-code-review/blob/main/README.md

Enhance review targeting by providing background context or requirements for the changes being reviewed.

```bash
ocr review --background "Adding rate limiting to the login API"
```

--------------------------------

### Select or Override Model for OCR Review

Source: https://github.com/alibaba/open-code-review/blob/main/README.md

Specify a particular LLM model for the review, either generally or for a specific commit.

```bash
ocr review --model claude-opus-4-6
ocr review --commit abc123 --model claude-sonnet-4-6
```

--------------------------------

### Configure LLM Manually

Source: https://github.com/alibaba/open-code-review/blob/main/README.md

Sets LLM configuration parameters directly via the CLI. Configuration is stored in ~/.opencodereview/config.json.

```bash
ocr config set llm.url https://api.anthropic.com/v1/messages
ocr config set llm.auth_token your-api-key-here
ocr config set llm.model claude-opus-4-6
ocr config set llm.use_anthropic true
```

--------------------------------

### Check Rule Application to a File

Source: https://github.com/alibaba/open-code-review/blob/main/README.md

Verify which custom review rule would apply to a specific file, or check a file against a specified rule.

```bash
ocr rules check src/main/java/com/example/Foo.java
ocr rules check --rule custom.json src/main/resources/mapper/UserMapper.xml
```

--------------------------------

### Perform Code Review (Branch Range)

Source: https://github.com/alibaba/open-code-review/blob/main/README.md

Compares two Git references (branches or tags) and reviews the changes between them.

```bash
# Branch range — compare two refs
ocr review --from main --to feature-branch
```

--------------------------------

### Add Custom Review Rules

Source: https://github.com/alibaba/open-code-review/blob/main/examples/github_actions/README.md

Use the `--rule` flag with the `ocr review` command to specify a custom JSON file containing your review rules. This allows for tailored code analysis.

```yaml
- name: Run OCR review
  run: ocr review --rule ./my-rules.json --from origin/${{ github.base_ref }} --to origin/${{ github.head_ref }}
```

--------------------------------

### Configure LLM using Persistent Config

Source: https://github.com/alibaba/open-code-review/blob/main/skills/open-code-review/SKILL.md

Sets LLM configuration persistently using the 'ocr config set' command. This includes URL, authentication token, model, and Anthropic usage.

```bash
ocr config set llm.url https://api.anthropic.com/v1/messages
ocr config set llm.auth_token <api-key>
ocr config set llm.model claude-opus-4-6
ocr config set llm.use_anthropic true
```

--------------------------------

### Review Working Copy

Source: https://github.com/alibaba/open-code-review/blob/main/plugins/open-code-review/skills/open-code-review/SKILL.md

Use this command to review staged, unstaged, and untracked changes in your working directory. The -b flag is used for providing business context.

```bash
ocr review --audience agent -b "context"
```

--------------------------------

### Review Specific Commit with Verbose JSON Output

Source: https://github.com/alibaba/open-code-review/blob/main/README.md

Perform a review on a specific commit, outputting the results in verbose JSON format and targeting an agent audience.

```bash
ocr review --commit abc123 --format json --audience agent
```

--------------------------------

### Rule File with Path Filtering

Source: https://github.com/alibaba/open-code-review/blob/main/README.md

Includes rules along with 'include' and 'exclude' fields to control which files are considered for review. 'Exclude' patterns take precedence over 'include'.

```json
{
  "rules": [
    {"path": "**/*.java", "rule": "Check for null safety"}
  ],
  "include": ["src/main/**/*.java", "lib/**/*.kt"],
  "exclude": ["**/generated/**", "vendor/**"]
}
```

--------------------------------

### Add Custom Review Rules

Source: https://github.com/alibaba/open-code-review/blob/main/examples/gitlab_ci/README.md

Execute OCR review with a custom rules JSON file. Ensure the rules file is accessible within the CI environment.

```yaml
script:
  - ocr review --rule ./my-rules.json --from origin/$CI_MERGE_REQUEST_TARGET_BRANCH_NAME --to $CI_COMMIT_SHA
```

--------------------------------

### Copy OCR Command File for Claude Code (Project-level)

Source: https://github.com/alibaba/open-code-review/blob/main/README.md

Command to copy the OCR command file to a project's .claude/commands directory for team-wide use.

```bash
mkdir -p .claude/commands
curl -o .claude/commands/open-code-review.md \
  https://raw.githubusercontent.com/alibaba/open-code-review/main/plugins/open-code-review/commands/review.md
```

--------------------------------

### Provide Background Context for Review

Source: https://github.com/alibaba/open-code-review/blob/main/examples/github_actions/README.md

Use the `--background` flag with the `ocr review` command to pass additional context, such as the PR title, to help OCR better understand the changes and provide more relevant comments.

```yaml
- name: Run OCR review
  run: ocr review --background "${{ github.event.pull_request.title }}" --from origin/${{ github.base_ref }} --to origin/${{ github.head_ref }}
```

--------------------------------

### Proper malloc/free Pairing in C

Source: https://github.com/alibaba/open-code-review/blob/main/internal/config/rules/rule_docs/c.md

Ensure every allocated memory block using malloc is freed using free. Set the pointer to NULL after freeing to prevent double-freeing.

```c
// Bad
char* buffer = malloc(1024);
// use buffer...
// forgot to free memory

// Good
char* buffer = malloc(1024);
if (buffer != NULL) {
    // use buffer...
    free(buffer);
    buffer = NULL;
}
```