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

# Linear

> File a Linear issue for every finding above your severity threshold, and keep the issue and the finding in step.

Enoki files an issue for each new finding above the severity threshold you
choose, in the Linear team you choose. The issue and the finding then stay in
step: complete the issue in Linear and the finding is marked fixed in Enoki;
verify its fix in Enoki and the issue is completed in Linear, or dismiss the
finding and the issue is canceled.

## Connect

Only **workspace admins** manage integrations. Connecting does not start
filing: the connection is stored without a team and switched off.

<Steps>
  <Step title="Open integration settings">
    Sign in at [app.enokilabs.ai](https://app.enokilabs.ai) and go to
    **Settings → Integrations**.
  </Step>

  <Step title="Connect Linear">
    Click **Connect Linear** and authorize the read and write access it asks
    for: read resolves your organization and its teams, write creates issues,
    comments on them, and moves them between states.
  </Step>

  <Step title="Choose a team and a threshold">
    Pick the Linear team to file into, and the severity threshold. A new
    connection starts at **High and above**; you can widen it to **All
    severities** or pick another band.
  </Step>

  <Step title="Turn it on">
    Click **Create issues automatically**. The next run that opens a
    qualifying finding files it.
  </Step>
</Steps>

## What gets filed

One issue per finding, per workspace. The title is the finding's name, as
its page in Enoki shows it; the body carries the category, strategy,
severity, description, expected behaviour, and a link to the finding in Enoki.

A run files every open finding it sees at or above your threshold, including
ones that existed before you connected, and a finding re-graded upward across
the threshold is filed at that point. To file one no run has seen, use **Push
to Linear** on the finding.

## How the two stay in step

| In Enoki | In Linear |
| - | - |
| A new finding above your threshold | A new issue in your chosen team |
| A fix verified by a replay (`fixed_verified`) | The issue is completed, commenting that the fix was verified |
| A finding dismissed | The issue is canceled, commenting that it was dismissed |
| A finding observed again after being fixed (`regressed`) | The issue is reopened, commenting that it was observed again |
| A fix that did not hold (`still_broken`) | The issue is reopened, commenting that the attack still works |

| In Linear | In Enoki |
| - | - |
| The issue moves to a completed state | The finding is marked **Fixed**, and its attacks are re-run |
| The issue is cancelled, or moved to any other state | Nothing. Only completion is read as a fix |

Completing an issue is you asserting the fix. Enoki then checks it: the
finding's attacks are replayed, and the verdict comes back to the issue. If
they no longer work the finding reaches **Fixed (verified)** and the issue
stays completed. If they still work the finding goes to **Still broken** and
the issue is reopened.

## Good to know

* **Turning issue creation off stops both directions.** Existing issues stay
  as they are.
* **The re-run is a replay, not an assessment.** It re-runs that finding's
  own attacks against your agent, spends none of your assessment allowance,
  and is capped per finding per day, so completing an issue repeatedly cannot
  queue repeated traffic.
* **A failed push is visible and can be retried.** The finding shows a
  **Linear push failed** badge. A run that sees the finding again retries it,
  and an admin can retry it at any time with **Retry push to Linear** on the
  finding. The Integrations tab shows when the bridge last delivered, and
  whether it failed.
* **Enoki only touches issues it created**, in the team you selected.
* **Disconnecting removes Enoki's access in Linear, and deletes the stored token
  and the record of which issue stands for which finding.** Issues already in
  Linear stay there, untracked;
  after a reconnect, a finding a run sees again gets a new issue.
* **Redacted findings are not filed in full.** If your plan withholds finding
  detail in the dashboard, it is withheld before an issue body is built.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.