Agents in Superconductor can now read and write to Linear while they work on your tickets.
If your team works in Linear, the description you paste into a Superconductor ticket is rarely the whole story. Perhaps the acceptance criteria got tightened in a comment, a sibling issue explained why the last attempt was reverted, or the actual priority lives in the current cycle. Linear MCP lets agents go get that context themselves.
Some workflows this opens up:
- Hand an agent an issue identifier and let it plan the implementation from the acceptance criteria and the discussion in that Linear issue
- Ask agents to check related and duplicate issues first, so they know which approaches your team already ruled out
- Have agents create new Linear issues for any follow-up tasks that surface during your code review in Superconductor, with repro steps, without leaving the chat
- Drop a comment on the Linear issue when an implementation is ready for a teammate to look at
- Ask for a plain-language summary of a cycle or project in Linear when you are deciding what to pick up next
The Linear MCP helps carry your context in both directions, so that PMs, designers, and engineers can seamlessly work between Linear and Superconductor without having to re-type anything.
To turn on the Linear MCP, go to Project Settings -> Skills and MCPs, enable Linear, and paste in a Linear API key. We connect through Linear's hosted MCP server, and the key is stored encrypted and handed to agents only while the integration is on. Since the key is shared by the project, give it the narrowest set of permissions your agents actually need.
Linear joins the rest of the tools agents can reach from a Superconductor session. Check out our full list of supported Skills and MCP servers.
Which MCP should we add next? Let us know!