Screen interview code submissions with a Codex review agent
A commenter uses Codex with a targeted agents.md for front-line code reviews, rejecting submissions with too many files changed or added.
- What they did
- A commenter said they set up ChatGPT Codex with a targeted agents.md file to handle first-pass reviews of coding submissions. The agent reviewed each submission against simple gate rules and automatically rejected ones with too many files changed or added; the human still chose the rules and used the agent as the front-line screener.
- How it went
- The post gives no metrics, examples, or comparison to manual review. It only claims this was being used to fight low-quality, AI-assisted submissions earlier in the hiring funnel.
- Worth knowing
- The useful catch is how narrow the filter was: they tuned it around obvious review heuristics like file-count and added/changed-file limits, not a full autonomous hiring decision.
Try it yourself with Codex
Write an agents.md for [repo or take-home project] that tells you how to screen candidate submissions: flag any submission that changes or adds more than [number] files, doesn't follow [requirement], or lacks tests. Then review the submission in [folder path] against it and give me a pass or reject recommendation with reasons. You're done when I have a short review I can use; don't contact the candidate or post anything anywhere.
Want help getting this working? Join the Vibe Coding Academy ↗ Run by Alex Finn, who also runs this site.
Source: Reddit · Undated
Related use cases
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.
Screen research papers against criteria
Codex reads a folder of PDFs to find articles that disagree with an argument, screens about fifty papers for a systematic review, and cleans messy CSV or Excel files.
Port Claude Code skills to Codex
The author moved 201 Claude Code skills to Codex in an afternoon using a skill inside Claude Code, and reports what transferred and what broke.
Build a C compiler with a team of parallel agents
Anthropic's engineering write-up on tasking Opus 4.6 agent teams with building a C compiler.
Delegate web research from Claude Code to Comet
An MCP server lets Claude Code hand web navigation, login-walled pages and research to Comet while it keeps coding, then returns results.
Manage NixOS server configs with Claude Code
The author uses Claude Code to maintain the NixOS config repo for their self-hosted machines, blocked from running nixos-rebuild switch and reviewing each change before applying it manually.
Get 5 like this every morning
The best things people got an AI agent to do, each with the prompt to try it.