backlog-mcp-server

What it does

This server integrates Claude with the Backlog project management API, exposing tools to create and manage issues, projects, milestones, wiki pages, and Git repositories. You can query existing data, post comments on issues, and perform CRUD operations on Backlog resources through Claude conversations, with support for GraphQL-style field selection to optimize response payloads.

Who it's for

Engineering teams and product managers using Backlog who want to automate issue management workflows, generate project reports, or integrate Backlog data into Claude-powered analysis. Useful for DevOps engineers managing shared Backlog instances via MCP's OAuth authentication flow.

Common use cases

• Create and update issues directly from Claude conversations, referencing project requirements.
• Query issue lists and project status for status reports or metrics.
• Post comments on issues and pull requests to automate code review engagement.
• Manage versions and milestones programmatically based on planning data.
• Access wiki pages and repository information for documentation or analysis.

Setup pitfalls

• Requires BACKLOG_DOMAIN and BACKLOG_API_KEY environment variables; no API key means no tools are available. For shared deployments, use OAuth 2.0 instead (requires separate Backlog OAuth app registration).
• The continuous integration suite is currently failing — validate functionality locally before relying on production deployments.
• HTTP transport mode exposes full Backlog API access; never bind to 0.0.0.0 without TLS and host verification unless behind a trusted proxy.
• Writes to the filesystem depending on usage patterns — ensure the server process has appropriate permissions.