Question 1

Can I assign a Linear issue to an AI coding agent?

Accepted Answer

Yes. Replicas shows up in Linear as an assignable agent. Assign or delegate an issue to it, or mention @Replicas in a comment, and it starts a coding agent in a cloud VM.

Question 2

What does the agent see from the issue?

Accepted Answer

The title, the description and its images, the parent issue, earlier comments, and your comment if you mentioned @Replicas.

Question 3

Does it update the issue status?

Accepted Answer

Yes. It moves the issue to In Progress when work starts, In Review when its turn ends with a PR open, and Blocked when it needs input or fails. You can map each of these to your team’s own state names.

Question 4

Which coding agent does Linear use?

Accepted Answer

The agent defaults of the person who triggered it when their Linear account is linked, then any Linear-specific defaults an admin set, then the organization defaults. Follow-up comments keep the workspace’s agent and model.

Question 5

What happens when I close the issue?

Accepted Answer

Marking it Done, Canceled or Duplicate archives the workspace. Its conversation history is kept and it can be restored.

You	Replicas
Assign or delegate an issue to Replicas	Starts a workspace; the human assignee keeps ownership
Mention @Replicas in a comment	Scopes the work to what your comment asks for
Watch the agent session	Streams progress, with workspace and PR links
Let it work	Moves the issue to In Progress, then In Review with a PR
Comment again	Wakes the same workspace and keeps the same agent and model
Start a comment with /plan	Returns a plan instead of code
Mark the issue Done or Canceled	Archives the workspace, keeping its history

Assign Linear issues to an AI coding agent.

From Linear to a reviewed pull request.

What the agent can do from Linear.

Connect Linear in 4 steps.

Linear questions