Let an agent work through your code changes overnight
A developer had Amp review open code changes, answer Codex review comments and run end-to-end checks overnight; he says it got confusing and ate almost all his usage.
Done with Amp
What they did
A developer talked through an approach and milestones with Amp for his open-source repo. He then let it run overnight. He asked it to review every pending change, wait for checks, fix Codex review comments and re-request review. He also asked for an end-to-end test proving everything worked.
How it went
He says Claude Code ran out of usage for five hours, so the work shifted to Amp's agents. He reports the many threads became confusing to follow and used almost all his plan's usage.
Worth knowing
Leaving many parallel agent threads unattended overnight can burn through a usage plan fast and leave you unsure what happened.
Try it yourself with Amp
Work through the open pull requests in [repo URL] one at a time. For each, wait for the automatic checks to finish, read any review comments, fix them, and ask the reviewer to look again. Keep a short running log of what you changed in each pull request. Stop if usage passes [usage limit]. Do not merge anything. At the end, run an end-to-end test with [example app]. Show me the log and the test results for approval.
Swap the highlighted [parts] for your own details.
Source: X · @GaurangKaria · Oct 17, 2026
Written with AI assistance from the original post and checked against our editorial guidelines. Results are as reported by the source, not independently verified. Spotted a mistake? Tell us.
