agentusecasesAll 965 use cases
Software & DevOps

Turn tickets into merged pull requests

Pulls tasks from GitHub Issues or Linear, then runs Claude Code or Codex in isolated Kubernetes pods to carry each one through to a PR.

Done withClaude CodeCodex

What they did
Optio is a self-hosted tool you run on Kubernetes. You assign it a GitHub Issue, Linear ticket or Jira card. It finds or creates a long-lived pod for the repo and adds a git worktree. It runs the agent there and opens a PR. A PR watcher polls CI and review state, and resumes the agent with context on CI failures, merge conflicts or review feedback. A review agent also runs.
How it went
The README says the pipeline runs end to end, squash-merging when CI is green and the review is done, then closing the linked issue and recording cost. It gives no success rates or benchmark numbers.
Worth knowing
It needs Kubernetes v1.33 or newer, Helm, Postgres and Redis. Auto-resumes are capped by the OPTIO_MAX_AUTO_RESUMES setting, so stuck PRs can still need a person.

Try it yourself with Codex

Look at the open issues in my repo [owner/repo] labeled [label, e.g. good-first-issue] and pick the [number, e.g. 3] smallest ones. For each, work on its own branch, make the fix, run the tests, and prepare a draft pull request description. Pause and show me each diff and PR description before you push anything or open a PR.

Read the original ↗

Discussion on HN · Mar 25, 2026

Five of these in your inbox every morning

The best things people got an AI agent to do, each with the prompt to try it.

More like this