OpenAI just announced that Codex is now generally available, and they're shipping some really practical features that make it way more useful for engineering teams.
What's New#
The big additions are a Slack integration, the Codex SDK, and better admin tools. But honestly, the SDK and GitHub Actions integration are what really caught my attention.
Slack Integration: Delegate Like a Coworker#
You can now tag Codex directly in Slack channels or threads to kick off tasks. Just treat it like another team member. Need someone to investigate a bug or refactor something? Drop it in Slack and let Codex handle it in the background.
Check out the demo from DevDay to see how smooth this workflow is.
The SDK: Embed Codex Anywhere#
This is where it gets interesting. The Codex SDK lets you embed the same agent into your own tools and workflows with just a few lines of TypeScript code.
import { Codex } from "@openai/codex-sdk";
const codex = new Codex();
const thread = codex.startThread();
const result = await thread.run(
"Make a plan to diagnose and fix the CI failures"
);
console.log(result);
You can continue on the same thread or resume past threads. It handles context management automatically, and outputs are structured so you can parse agent responses easily.
GitHub Actions: Auto-Fix Your CI#
Here's the really cool part. You can now add Codex directly into your CI pipeline to automatically fix failing tests. When your CI fails, Codex spins up, diagnoses the issue, applies a minimal fix, and opens a PR for review.
The setup is straightforward. Add this to .github/workflows/codex-autofix.yml:
name: Codex Auto-Fix on Failure
on:
workflow_run:
workflows: ["CI"]
types: [completed]
permissions:
contents: write
pull-requests: write
jobs:
auto-fix:
if: ${{ github.event.workflow_run.conclusion == 'failure' }}
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v4
- uses: openai/codex-action@main
with:
openai_api_key: ${{ secrets.OPENAI_API_KEY }}
prompt: >-
Read the repo, run tests, identify minimal fix,
implement only that change. Keep it surgical.
That's it. Codex watches for failures, fixes them, and submits a PR. You review and merge like any other contribution.
The full setup guide is here on the OpenAI docs.
Bottom Line#
The GA release makes Codex feel less like an experiment and more like a production tool. The SDK and GitHub Actions integration are the standout features here. Being able to automate CI fixes and embed Codex into your own workflows opens up a lot of possibilities.