How to connect Sourcegraph to Claude Cowork

Cowork is Anthropic's AI agent for knowledge work. Think of it as Claude Code for everything else. It works autonomously with your computer, local files, and applications to accomplish complex tasks.

This guide walks you through the easiest and most secure way to connect your Sourcegraph account to Cowork via Composio Connect, enabling it to show all languages used in repo backend-service, compare latest two commits in project web-ui, list all files under src/utils in repo app-core, and more such actions on your behalf without compromising your account security.

Connecting Sourcegraph to Claude Cowork

1. Open Customize

In Claude Desktop, click Customize in the left sidebar, then select Connectors and click the + icon at the top.

2. Add the Composio MCP server

Click Add custom connector and paste in the Composio MCP server URL:

Claude Cowork is Anthropic's agent for general knowledge work. It can use your computer, files, and connected applications to complete longer-running tasks across your work tools.

With Composio Connect, Cowork can securely access apps like Sourcegraph through MCP without you sharing account credentials directly with the agent.

The Sourcegraph MCP server is an implementation of the Model Context Protocol that connects your AI agent and assistants like Claude, Cursor, etc directly to your Sourcegraph account. It provides structured and secure access to your codebase, so your agent can perform actions like searching repositories, inspecting commits, fetching file contents, and analyzing code languages on your behalf.

• Repository discovery and listing: Let your agent list all repositories in your Sourcegraph instance, making it easy to navigate and select codebases for deeper analysis.
• Commit inspection and comparison: Have the agent fetch details about specific commits or compare two commits to see file-level changes and diffs in any repository.
• File content retrieval: Ask your agent to fetch the raw contents of any file in your codebase—no cloning or manual browsing required.
• Repository file and language enumeration: Enable the agent to list all files, directories, and programming languages used in a given repository for a comprehensive code overview.
• User identity and permissions checks: Let your agent verify the authenticated user or check site settings permissions before performing sensitive actions.