On August 12, a developer going by MPIsaac-Per dropped a project on Hacker News that solves a real problem for power users living between multiple tools. The project, called linear-claude-bridge, creates a direct connection between Linear's agent framework and Claude Code running locally on your machine.

How the Bridge Works

The setup leverages Linear's built-in OAuth app functionality to register Claude Code as an official agent. Once configured, you can either mention the agent directly in a ticket comment or delegate an entire issue to it. The bridge then fires up a Claude Code session against whatever knowledge base you're working with—typically Obsidian for users running this kind of setup—and returns insights, code snippets, or analysis back into the Linear thread where your team can see them without leaving their project management tool.

Why Local-First Matters

Here's what makes this interesting: everything runs on your hardware. No API calls to external services beyond Claude itself, no data escaping your environment. For developers who keep sensitive architecture docs in Obsidian or maintain internal knowledge bases they don't want indexed elsewhere, this bridges the gap between private tooling and collaborative platforms without compromising the local-first ethos that many in the hacker community still care about deeply.

The Developer Stack Behind This

The author describes their daily workflow as keeping everything locally in Obsidian while running Claude Code against it throughout the day. Linear handles project tracking. Before this bridge existed, bouncing between these tools meant context-switching or manually copying information back and forth. Now the knowledge base queries happen inline with ticket discussions—where the work is actually tracked.

Key Takeaways

  • Uses Linear's native OAuth agent registration to avoid custom webhook hacks
  • Claude Code sessions run entirely local—no intermediary services required
  • Directly surfaces Obsidian (or other local) knowledge into collaborative ticket context
  • Open source implementation available on GitHub for customization and auditing
  • Demonstrates growing pattern of bridging personal AI tooling with team-oriented platforms

The Bottom Line

This is the kind of glue code that makes toolchains actually usable rather than just technically functional. MPIsaac-Per identified a workflow friction point and solved it with minimal infrastructure. If you're running Claude Code locally against private knowledge bases while managing work in Linear, this is exactly the missing piece you've probably already tried to jury-rig differently.