# my claude limit ran out mid week so codex ran my repo for 4 days. notes from taking it back — retained record

> Retained reddit record from reddit:ClaudeCode.

Canonical: https://highsignal.app/data/records/5a989d9d91911033

[← sources](https://highsignal.app/data)


source: [reddit](https://highsignal.app/data/reddit) (reddit:ClaudeCode)  published 2026-11-11 06:55 ingested 2026-11-12 02:42

[original source ↗](https://www.reddit.com/r/ClaudeCode/comments/1xtb0v4/my_claude_limit_ran_out_mid_week_so_codex_ran_my/)

## retained text



```
solo dev, one repo, claude code does most of the work. last week i burned through my usage and instead of waiting i handed the whole thing to codex with a written brief. both read the same [AGENTS.md](http://AGENTS.md) so i figured it would mostly just work.

it kind of did? codex merged about 15 PRs in 4 days. the queue i left got done and main never broke badly.

the part i didnt expect was coming back. it had left 4 draft PRs, three of them stacked on each other, and when claude picked them up every one needed review rounds before i trusted it. one took 4 rounds. there was also a test on main that had started failing because it depended on the date, and nobody noticed for a day lol.

the two also write rules very differently. on a side task codex kept adding a new rules doc on top of the old one until the newest contradicted the first, and then it followed whichever it read last. when claude came back the first thing it did was collapse them into one file.

what id do differently next time is a much shorter brief, and one handover file thats the only thing that has to stay true.

anyone else swapping agents mid project when the limit hits? what broke for you on the way back?
```



document fetched 2026-11-12 00:43 · canonical https://reddit.com/r/ClaudeCode/comments/1xtb0v4/my_claude_limit_ran_out_mid_week_so_codex_ran_my
