The Omegro AI Playbook plugin

A plugin for Claude Code and Codex. It diagnoses what slows your team down and gives you skills to fix it, inside your repository, against your actual code.

Get started

Works with Claude Code and Codex. Install and authenticate either one first.

Enter your GitHub username to get access to the plugin repository.

GitHub username

Accept the invitation, then run these two commands in Claude Code or Codex:

# add the Omegro playbook marketplace (one time)
/plugin marketplace add hom-omegro/omegro-playbook

# install the QA consult plugin
/plugin install omegro-playbook

Run both commands inside Claude Code or Codex, not in a regular terminal.

Entered the wrong username? Request again

How the consult works

The session takes 15 to 20 minutes. You talk to it in plain English.

1. Quick checks
It confirms this is the right repo and asks whether you can read the code yourself or are asking on behalf of a team.
2. Frame check
Two questions about what you are trying to achieve and what you measure today. If your problem is not really a QA problem, it says so and points you to the CoE instead of running the wrong consult.
3. Scoping questions
Up to four short questions, one at a time, about how your team verifies changes, what QA checks by hand, what you tried before, and where a silent break would hurt most. Answer specifically: "the month-end billing run" beats "the usual stuff".
4. Evidence check (you choose)
It asks before it reads anything. If you say yes, it looks at your test folders and CI config only, read-only, never your application code. If you say no, it asks a few more questions instead.
5. Diagnosis and plan
It names one diagnosis, ranks your actual flows, gives you a concrete first test, and writes a plan with one next step, a check date, and a named owner. The plan goes to docs/ai-consult/ in your repo. No code, credentials, or customer details.
6. Send it back
Type /export, save the session, and email it plus the plan to the CoE with three quick answers: did the diagnosis match, would you do the next action, and what felt off.

The consult writes one file and sends nothing. It will not change anything in your repo without telling you.

What you get

Consults

Run /playbook-consult to start.

Skills

Separate from the consult. Your agent loads each automatically at the moment of need. For example, when you start the first action in your QA plan, the agent loads the right skill without you typing anything.