Sentry Auto Fix for Cursor

Integrate Sentry directly into Cursor/VSCode. Works with both Sentry.io (cloud) and self-hosted instances. View issues in a sidebar panel and generate AI-powered fix prompts with a single click.

Features

🔌 Connect to Sentry - Works with sentry.io or self-hosted instances
🎨 Modern Card UI - Beautiful card-based interface with visual hierarchy
🔍 Search & Filter - Instantly search issues and filter by project with checkboxes
📋 Browse Issues - View all unresolved issues from your projects in a dedicated sidebar
🤖 AI-Powered Fixes - Generate comprehensive prompts for Cursor AI with one click
🔄 Real-time Sync - Refresh issues on demand to stay up-to-date
📊 Issue Details - See event counts, affected users, and timestamps at a glance
🛠️ MCP Server - Let AI agents search, analyze, and resolve issues via Model Context Protocol
✅ Resolve Issues - Mark issues as "resolved in next release" directly from the UI

Installation

Install the extension in Cursor/VSCode
Configure your Sentry connection (see Configuration below)
The Sentry Issues panel will appear in the Activity Bar (sidebar)

Configuration

Open your settings (Cmd/Ctrl + ,) and search for "Sentry Auto Fix". Configure the following:

Required Settings

Sentry URL: Your Sentry instance URL
- Cloud (sentry.io): https://sentry.io ← Most common
- Self-hosted: https://sentry.example.com
- Do not include trailing slash
API Token: Your Sentry authentication token
- Create one at:
  - Cloud: https://sentry.io/settings/account/api/auth-tokens/
  - Self-hosted: https://your-sentry-url/settings/account/api/auth-tokens/
- Required scopes: project:read, event:read, event:write
Project Slugs: Array of projects to monitor
- Format: ["organization-slug/project-slug"]
- Example: ["my-company/web-app", "my-company/api"]

Example Configuration (settings.json)

For Sentry.io (Cloud):

{
  "sentryAutoFix.sentryUrl": "https://sentry.io",
  "sentryAutoFix.apiToken": "sntrys_your_token_here",
  "sentryAutoFix.projectSlugs": ["my-org/frontend", "my-org/backend"]
}

For Self-Hosted:

{
  "sentryAutoFix.sentryUrl": "https://sentry.example.com",
  "sentryAutoFix.apiToken": "your_sentry_api_token_here",
  "sentryAutoFix.projectSlugs": ["my-org/frontend", "my-org/backend"]
}

Usage

Viewing Issues

Click the 🐛 bug icon in the Activity Bar (left sidebar)
Issues display as cards grouped by project
Use the search bar at the top to filter issues
Click "📁 Projects" dropdown to show/hide specific projects
Click "🔄 Refresh" to reload issues from Sentry

Fixing Issues with AI

Find an issue in the sidebar
Click the "🤖 Fix with AI" button on the issue card

The extension will:

Fetch the latest event details from Sentry
Generate a comprehensive prompt including:
- Issue title and error message
- Full stack trace with file locations
- User actions (breadcrumbs) leading to the error
- Environment tags (browser, OS, runtime, etc.)
Send the prompt to Cursor AI chat (or copy to clipboard)

Resolving Issues

Click the "✅ Resolve in Next Release" button to mark an issue as resolved in Sentry. The issue will disappear from the list after the next refresh.

Opening Issues in Sentry

Click the "🌐 Open in Sentry" button on any issue card to view full details in your browser.

MCP Server

The extension includes an MCP (Model Context Protocol) server that allows AI agents to interact with your Sentry issues.

Setup

The MCP server is automatically registered when you activate the extension. After configuring your Sentry credentials, you'll see a message confirming registration.

Important: After the MCP server registers, you must fully restart Cursor (quit and reopen) for the AI to recognize the new tools.

{
  "mcpServers": {
    "sentry": {
      "command": "node",
      "args": ["/Users/ruben/Projects/sentry-auto-pr-mcp/dist/mcp-entry.js"],
      "env": {
        "SENTRY_URL": "https://your-sentry-url.com",
        "SENTRY_API_TOKEN": "your_token_here",
        "SENTRY_PROJECT_SLUGS": "org/project1,org/project2"
      }
    }
  }
}

After configuring, restart Cursor completely (not just reload window).

Available Tools

The AI agent can now:

sentry_search_issues - Search/filter issues by keyword, project, or limit
sentry_get_issue_details - Get full details including stack traces and breadcrumbs
sentry_list_projects - List all configured projects
sentry_resolve_issue - Mark issue as resolved in next release

Example Usage

In Cursor AI chat, you can say:

"Show me all Sentry errors from the last week"
"Get details for issue MYAPP-123"
"Which project has the most issues?"
"Great, let's resolve issue MYAPP-123" ← Marks it as resolved!

Generated Prompt Format

The AI fix prompt includes:

Issue Summary: ID, title, severity, event count, affected users
Error Message: The actual error that occurred
Stack Trace: Complete call stack with file paths and line numbers
Affected Files: List of application files involved in the error
User Actions: Breadcrumbs showing what the user did before the error
Environment Tags: Browser, OS, runtime version, URL, etc.
Task: Specific instructions for the AI to analyze and fix

Troubleshooting

"Sentry Not Configured" Warning

Make sure you've set all three required settings:

sentryAutoFix.sentryUrl
sentryAutoFix.apiToken
sentryAutoFix.projectSlugs

"Failed to load Sentry issues"

Check that:

Your Sentry URL is correct (no trailing slash)
Your API token is valid and has the required scopes
Your project slugs are in the correct format: "org/project"
You have network access to your Sentry instance

API Token Scopes

Your Sentry API token needs these scopes:

project:read - To fetch project information and issues
event:read - To fetch event details
event:write - To resolve issues

Create a token at:

Cloud: https://sentry.io/settings/account/api/auth-tokens/
Self-hosted: https://your-sentry-url/settings/account/api/auth-tokens/

Commands

Sentry: Refresh Issues - Reload issues from Sentry
Sentry: Fix this issue using AI - Generate AI fix prompt for selected issue
Sentry: Open Issue in Sentry - Open issue in browser

Development

To build and run locally:

# Install dependencies
npm install

# Compile the extension
npm run compile

# Or watch for changes (recommended)
npm run watch

Then press F5 in VSCode and select "Run Extension (Watch Mode)" to launch with hot reload. When you make changes, press Cmd+R (Mac) or Ctrl+R (Windows/Linux) in the Extension Development Host window to reload.

See DEVELOPMENT.md for detailed development guide.

Requirements

VSCode/Cursor version 1.85.0 or higher
A Sentry account (sentry.io or self-hosted)
Valid Sentry API token with required scopes

Release Notes

0.0.1

Initial release:

Connect to Sentry.io (cloud) or self-hosted instances
View unresolved issues in a modern card-based UI
Search and filter issues by project
Generate AI fix prompts with full context (stack traces, breadcrumbs, tags)
Resolve issues directly from the extension
Open issues in browser
MCP server for AI agent integration

License

MIT

Contributing

Contributions are welcome! Please feel free to submit issues or pull requests.

Sentry Auto Fix

Ruben Seyferth