Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/jonwiggins/optio/llms.txt

Use this file to discover all available pages before exploring further.

Most AI coding tools stop after opening a pull request. Optio keeps going. Once the agent opens a PR, Optio monitors CI, watches for review feedback, handles merge conflicts, and merges the PR automatically when everything passes. This is the feedback loop — and it’s what turns Optio from a code generator into a workflow orchestrator.

The happy path

1

Task runs, PR is opened

The coding agent completes its work and opens a pull request. The task transitions to pr_opened.
2

PR watcher begins polling

Every 30 seconds, Optio polls the GitHub API for CI check status, review state, and merge readiness.
3

Code review agent runs (if enabled)

If code review is enabled for the repository, a review agent launches as a subtask — either immediately on PR open or after CI passes, depending on your configuration.
4

CI passes and review is approved

When all CI checks pass and any required review is approved (or review is not required), Optio triggers an auto-merge.
5

PR is squash-merged, issue closed

The PR is squash-merged into the default branch. If the task was linked to a GitHub Issue or Linear ticket, that ticket is closed automatically. The task transitions to completed.

Feedback scenarios

Things rarely go perfectly on the first try. Here is how Optio handles the three most common interruptions:
When CI checks fail, the PR watcher detects the failure and transitions the task to needs_attention. If auto-resume is enabled for the repository, Optio immediately re-queues the agent with the full CI failure output as context.The agent reads the failure output — test errors, lint warnings, build logs — and pushes a fix to the same branch. Once the fix is pushed, CI re-runs automatically, and the loop continues.
The number of automatic resume cycles is capped by the Max auto-resumes setting on the repository. The default is 10. After the cap is reached, the task stays in needs_attention for manual review.

Auto-merge

Auto-merge squash-merges the PR when all of the following are true:
  • All CI checks have passed
  • Any review subtask has completed (if review is enabled)
  • No reviewer has an outstanding “changes requested” review
  • The PR is still open
When the PR is merged, Optio closes any linked GitHub Issue or Linear ticket and marks the task completed.
Auto-merge is disabled by default. Enable it per repository in Repositories → [repo] → Settings → Auto-merge. Make sure your repository’s branch protection rules require CI to pass before merge, otherwise PRs could be merged with failing checks.

Per-repository settings

Both auto-resume and auto-merge are configured per repository. Navigate to Repositories, open a repository, and go to its Settings tab.
SettingDefaultDescription
Auto-resumeOffResume the agent automatically on CI failure, merge conflicts, and review feedback.
Max auto-resumes10Maximum number of automatic resume cycles per task before stopping.
Auto-mergeOffSquash-merge the PR automatically when CI passes and review is approved.
Auto-resume on reviewOffResume the agent with reviewer comments when changes are requested.
Disabling auto-resume does not stop the PR watcher. Optio will still track CI and review status and update the task state — it just won’t re-queue the agent automatically. You can always resume a needs_attention task manually from the task detail page.

Build docs developers (and LLMs) love