Codex Cli freezes executing commands
Resolved 💬 5 comments Opened Jan 10, 2026 by tradingfuturo Closed Jan 17, 2026
💡 Likely answer: A maintainer (github-actions[bot], contributor)
responded on this thread — see the highlighted reply below.
What version of Codex is running?
v0.80.0
What subscription do you have?
Business
Which model were you using?
gpt-5.2 codex medium
What platform is your computer?
Windows 11
What issue are you seeing?
Immediately freezes when trying to execute command. Stays looking as working but no advance. Is not a complete freeze, the interface works but no advances in tasks at all. Appears as if working but blocked internally.
https://github.com/openai/codex/issues/new?template=2-bug-report.yml&steps=Uploaded%20thread:%20019ba867-4b37-74b3-
a1d5-3f29aaa9e878
What steps can reproduce the bug?
Open and execute a task.
What is the expected behavior?
Advance normally as in the previous version
Additional information
_No response_
5 Comments
Potential duplicates detected. Please review them and close your issue if it is a duplicate.
Powered by Codex Action
Thanks for uploading your session details.
The problem is that the agent is running a
dotnet testcommand that never terminates, and this blocks further work.Ok but this is normal during development, and AI (or anyone) can make mistakes that can cause freezes specially with multithreading and incorrect locking patterns. What about the timeout? I usually improve tests adding timeouts to tests themselves (or explicitly tell AI to add them), but Codex should timeout with any jammed command.
I am having a similar issue.. It's just stuck. I am unable to press the submit button or hit yes/no in the menu dialogue.. This has happened every single time I use Codex since I installed it.. Using 5.3 Codex. CBA with this anymore.
<img width="515" height="674" alt="Image" src="https://github.com/user-attachments/assets/ec285442-c8c2-4d18-9fbf-64afdba10e62" />
So far my tests are not freezing at all since I am in a project maturity level where new tests are barely added anymore (80% line coverage and 70% branch). Also, I split the tests into smaller sets so they can run faster and I have less interference. But this is a latent issue. The agent should not freeze and have some timeout. I have found that LLMs are quite bad at estimating them but still a hard limit could be added so this doesn't cause a permanent freeze.