lgtm

KO | EN

gitlite — search

#python #hacktoberfest #react #ai #typescript #android #deep-learning #nodejs #java #windows #awesome #pytorch #ios #open-source

(★ 30)

A Visual Studio Code extension for reviewing pull requests or changes in git using GitHub Copilot Chat

lgtm Latest Version Download

Download Latest Version (.zip)

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

# LGTM - AI Code Review

[![Latest Release](https://flat.badgen.net/github/release/cpulvermacher/lgtm)](https://github.com/cpulvermacher/lgtm/releases)
[![License](https://flat.badgen.net/github/license/cpulvermacher/lgtm)](https://raw.githubusercontent.com/cpulvermacher/lgtm/master/LICENSE)

Use GitHub Copilot Chat to review pull requests or source code changes in Git, and catch bugs, areas for improvement, and other issues before merging.
Open the command palette and select `LGTM: Start Code Review...` to select what to review.

![Demo](https://raw.githubusercontent.com/cpulvermacher/lgtm/master/images/demo.gif)

## Getting Started

After starting a review with `LGTM: Start Code Review...`, LGTM will print review comments grouped by file and sorted by severity in the Chat sidebar.
Click on line numbers to jump to the location (if checked out).

### Reviewing Pull Requests

If you have the [GitHub Pull Requests](https://marketplace.visualstudio.com/items?itemName=GitHub.vscode-pull-request-github) extension installed, you can start a review by hovering over a pull request and clicking the LGTM icon:

![Reviewing a GitHub pull requests](https://raw.githubusercontent.com/cpulvermacher/lgtm/master/images/github_pr.png)

If you have the [Atlassian: Jira, Rovo Dev, Bitbucket](https://marketplace.visualstudio.com/items?itemName=Atlassian.atlascode) extension installed, you can also start a review for Bitbucket pull requests directly:

![Reviewing a Bitbucket pull request](https://raw.githubusercontent.com/cpulvermacher/lgtm/master/images/bitbucket_pr.png)

### Starting a Review From Chat

- Type `/review` followed by Return to bring up the same interactive menu as the `LGTM: Start Code Review...` menu command.
- You can specify git refs as arguments, e.g. `/review develop main`, or omit arguments to select refs interactively.

### Use in Agent Mode

LGTM is also available in agent mode, so you can include it as part of your worklflow.
For example, you might ask the agent to `Review the current changes using #reviewStaged and fix any severe issues.`
Assuming you have staged changes, the agent will start a review using LGTM and then act on the review comments. Consider specifying the severity of issues to fix, e.g. `... and fix any issues with severity >= 3.`

## Programmatic Use

Extensions and VS Code automation can call the hidden `lgtm.reviewChanges` command directly. It runs the same review flow in the background with a cancellable progress notification, then resolves with structured review comments, per-model results, and any errors.
Consumers that need to verify the command API can check the shared types in `src/types/ReviewChangesCommand.ts`.

```ts
// Review staged changes using the configured `lgtm.chatModel`.
const staged = await vscode.commands.executeCommand('lgtm.reviewChanges', 'staged');

// Review unstaged changes with one explicit review provider.
const unstaged = await vscode.commands.executeCommand(
    'lgtm.reviewChanges',
    'unstaged',
    'copilot:gpt-4.1'
);

// Review a topic/base pair with the configured default provider plus all configured preferred providers.
const branchReview = await vscode.commands.executeCommand(
    'lgtm.reviewChanges',
    'feature-branch',
    'main',
    'preferred'
);

// Use an options object when that is clearer for typed integrations.
const optionsReview = await vscode.commands.executeCommand('lgtm.reviewChanges', {
    topic: 'feature-branch',
    base: 'origin/main',
    models: ['copilot:gpt-4.1', 'copilot-code-review'],
});

// Equivalent staged/unstaged object forms are also supported.
await vscode.commands.executeCommand('lgtm.reviewChanges', { staged: true });
await vscode.commands.executeCommand('lgtm.reviewChanges', {
    scope: 'unstaged',
    models: 'preferred',
});
```

## Features

- **Only Copilot Required**: Defaults to using Copilot Chat for reviewing changes. Both free and paid plans are supported.
- **Model Selection**: Choose from language models available to VS Code, including Copilot Code Review for the same review experience as assigning Copilot to a pull request. Use the **LGTM: Select Chat Model** command in the Command Palette (`Cmd+Shift+P` / `Ctrl+Shift+P`) to change it.
- **Pull Request Integration**: Review pull requests from the GitHub Pull Request or Bitbucket extensions.
- **Automatic Fixes**: Use the `Fix` action on a review comment to fix it automatically. This will use the language model configured for inline chat.
- **Chat Integration**: Review content remains in chat history for follow-up questions by omitting `@lgtm`.
- **Project-Specific Context Files**: By default, `AGENTS.md` is added to the review prompt. Can be configured using the `Lgtm: Context Files` setting.
- **Custom Instructions**: Add global custom instructions via the `Lgtm: Custom Prompt` setting (e.g., change the language of review comments by adding `- In the final JSON output, use Spanish for the 'comment' field.`).
- **Agent Support**: Adds tools to enable automatic reviews in agent mode:
  - `#review`: Reviews changes between two git references (branches, tags, or commits)
  - `#reviewStaged`: Reviews only staged changes in your working directory
  - `#reviewUnstaged`: Reviews only unstaged changes in your working directory
  - Example usage: `After your changes, run all tests and run #reviewUnstaged to check your work.`

## Limitations

- Chat models listed under "Unsupported Models" can be selected but reviews will fail at runtime.
- Non-Copilot models (e.g. those added by directly adding an API key) may fail to work. For Claude language models, try the [Claude Model Provider](https://marketplace.visualstudio.com/items?itemName=cpulvermacher.claude-model-provider) extension.
- Large change sets may trigger chat model rate limits. Please wait before retrying.

## Data Usage

Source code changes and commit messages selected for review are sent to the chat model configured in the extension settings.

## Contributing

Contributions are welcome! If you have ideas, bug reports, or want to help improve LGTM, please open an issue or submit a pull request on [GitHub](https://github.com/cpulvermacher/lgtm).

lgtm

lgtm Latest Version Download

Today's picks

Was this content helpful?

Recent Feedback

lgtm

lgtm Latest Version Download

Today's picks

Was this content helpful?

Recent Feedback

Download README