Code from one AI is hard to trust blind. This developer runs a Docker harness where Claude writes and Codex reviews, works in it most of the day, and open-sourced it as moarcode.
Result Open-sourced as moarcode
“It has rough edges for sure, but it means I trust the code coming out much more than I did just Claude.”
Spotted: found in the wild. Details are as reported by the source above and have not been independently verified.
Wrote a Dockerfile and instructions so that Claude writes code while Codex reviews it, and published the setup as an open-source project called moarcode.
The developer spends most of the day in this harness. It has rough edges, but they trust the code coming out of it much more than code from Claude alone.
You are the reviewer. Another agent wrote the change below. Review it as a strict senior engineer: check correctness, edge cases, error handling, tests and security. List every problem with its file and line, mark each one blocking or minor, and propose the exact fix. Don't rewrite the whole change. Change: [paste diff]
A starting prompt written by Agents in the Wild. Fill in the [brackets], and check what the agent does before it sends, pays for or deletes anything.