FeaturedDev & IT Ops#Codex#GitHub#Draft PR#DevOps#Slack

Codex PR Fix Agent

Send a coding task to Codex, open a draft PR when it succeeds, and notify Slack when Codex needs follow-up input.

Workflow at a glance

The full canvas, before you import it

Click any node to see its config.

#Codex#GitHub#Draft PR#DevOps#Slack

Click a node to select it — same as the Heym editor; the panel shows its settings.

5 nodes · Free & source-available

Codex PR Fix Agent

This template wires the Codex node into a production-style coding workflow. A task prompt enters the workflow, Codex clones the configured GitHub repository inside Heym, applies code changes, and either returns a diff/draft PR or pauses with a follow-up question.

What this workflow does

  1. CodeTask receives the coding request
  2. CodexFix runs the local Codex CLI against the configured repository
  3. FixResult returns summary, validation, changed files, and draft PR URL
  4. NotifyCodexQuestion sends the Codex follow-up URL to Slack from the question branch

Use cases

  • Fix failing tests from an alert or issue triage workflow
  • Produce draft PRs from plain-English engineering tasks
  • Route missing requirements to Slack while the main workflow waits

Setup

Create an OpenAI Codex credential with a ChatGPT/Codex access token and a GitHub credential with repository access. In CodexFix, set the repository URL, base branch, setup command, and publish mode. In NotifyCodexQuestion, connect a Slack credential and channel.

Notes

Codex credentials use access_token, not an OpenAI API key, and are not exposed through $credentials. The template uses draft_pr mode; switch to diff_only if you only want the patch in the workflow output.

How to import this template

  1. 1Click Import → Copy JSON on this page.
  2. 2Open your Heym and navigate to a workflow canvas.
  3. 3PressCmd+V/Ctrl+V— nodes appear instantly.
  4. 4Add your API keys in the node config panels and click Run.
More workflow templates
View all templates
Heym
incident analysis · production AI
Observed across 100s of AI rollouts

AI workflows don't fail because of prompts.
They fail because of orchestration.

symptom · glue code01
5 tools
Scripts, vector DB, approval bot, tracing, browser runner — none of them talk.
symptom · visibility02
~0%
Observable behavior across the stack. Debugging is guesswork.
with heym · one runtime
1 canvas
Agents, RAG, HITL, MCP, traces & evals. Self-hosted. Observable.
AI-Native RuntimeProduction-Grade
github.com/heymrun/heym