> ## Documentation Index
> Fetch the complete documentation index at: https://docs.leme.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Linear

> Let agents query, create, and update Linear issues on your team's behalf

Connect Linear and your agents become fluent in your team's work: what's in progress, what's blocked, who owns what. They can also do the tedious part — filing well-written issues, updating status, leaving comments — with a human approving changes by default.

## What agents can do

| Capability                        | Details                                                             |
| --------------------------------- | ------------------------------------------------------------------- |
| List teams, members, and projects | The building blocks for any question about work.                    |
| List workflow states              | Your actual statuses — Backlog, In Progress, Done, and custom ones. |
| List and read issues              | Filterable by team; full detail by ID or key like `ENG-123`.        |
| Create issues                     | Title, Markdown description, assignee, status, project.             |
| Update issues                     | Change title, description, status, assignee, priority, or project.  |
| Comment on issues                 | Posted as the Leme app, visible to the whole team.                  |

Reading runs freely by default; creating, updating, and commenting ask for approval. See [Permissions](#permissions).

## Try asking

* *"What is the mobile team working on right now?"*
* *"List everything assigned to Ana that's still open."*
* *"Summarize ENG-123 and its current status."*
* *"Create an issue for the login timeout bug we just discussed, assign it to the platform team."* — created after approval.
* *"Move ENG-88 to In Review and add a comment linking this conversation's summary."* — applied after approval.

## Connect Linear

<Steps>
  <Step title="Open Integrations">
    In the Leme dashboard, go to **Integrations** and select **Linear**.
  </Step>

  <Step title="Authorize in Linear">
    Select **Connect** and approve on Linear's consent screen. You need permission to install applications in the Linear workspace.
  </Step>

  <Step title="Enable it on your agents">
    Attach the connection to the agents that should use it and review **Agent permissions**.
  </Step>
</Steps>

Linear is a **shared** connection: one authorization serves the workspace, and agents act as the Leme app — issues and comments they create are attributed to it, so it's always clear what an agent did.

## Permissions

| Action                                | Category               | Default      |
| ------------------------------------- | ---------------------- | ------------ |
| List and read teams, projects, issues | Read                   | Allow        |
| Create and update issues              | Write                  | Ask approval |
| Comment on issues                     | External communication | Ask approval |

Adjust in **Integrations → Linear → Agent permissions** — see [Permissions & approvals](/permissions). Teams that let agents triage often allow issue creation while keeping updates and comments behind approval.

## Troubleshooting

<AccordionGroup>
  <Accordion title="The agent can't find an issue by its key">
    Issue keys are team-scoped (`ENG-123`). If the key is right, the Leme app may not have access to that team's issues — check the app's access in Linear's workspace settings.
  </Accordion>

  <Accordion title="Creating an issue fails with an invalid field">
    Assignees, statuses, and projects are set by ID, and the agent resolves names to IDs first. If a name is ambiguous ("two Anas"), tell the agent which one you mean.
  </Accordion>

  <Accordion title="The connection shows as needing attention">
    The authorization expired or the app was revoked in Linear. Reconnect from **Integrations → Linear**; permissions and agent assignments are preserved.
  </Accordion>
</AccordionGroup>