Use Codex as a first-pass code submission filter

A commenter used Codex with a custom agents.md to check candidate code and automatically turn away submissions that broke simple file-change limits.

Done with CodexChatGPTยท Beginner

What they did
They configured ChatGPT Codex with a tailored agents.md so it could handle the initial review of coding take-home work. The agent compared each submission with a short list of rules and flagged it for rejection if the number of files changed or added was too high. The person behind the setup defined the thresholds and treated the agent as an intake screen, not the final judge.
How it went
They used it to catch weak, AI-generated-looking submissions earlier in the hiring flow. The post did not include any counts, examples, or evidence that it worked better than a manual review.
Tips
A small rule set is easier to trust than a broad autonomous review. Keep a person in charge of the final call, especially when the agent is only checking obvious signals like file counts or test coverage.

Try it yourself with Codex

Write an agents.md for [repository or take-home project] that tells Codex how to triage candidate submissions in [folder path]. Add simple checks such as rejecting work that changes or adds more than [number] files, misses [requirement], or lacks tests. Review the submission and give me a brief pass/reject recommendation with reasons. Pause and ask for my approval before any irreversible action, and do not message the candidate, submit anything, or modify files on your own.

Fill in the [brackets]. Agents can make mistakes, so review anything before it is sent, paid or deleted.

Read the original on reddit.com โ†—

Get 3 agent uses like this every morning

With the day's top AI news. Free, and one click to unsubscribe.