run_in_container.sh forces --full-auto and discards user flags for approval-mode
Resolved 💬 1 comment Opened May 2, 2025 by braun-viathan Closed Aug 7, 2025
What version of Codex is running?
0.1.4161210
Which model were you using?
o4-mini
What platform is your computer?
Linux 6.8.0-58-generic x86_64 x86_64
What steps can reproduce the bug?
- Run the script with any command line option like:
--approval-mode suggest. - Observe that Codex still starts in Full Auto.
What is the expected behavior?
– the script should use the provides approval-mode as the model param
What do you see instead?
– --full-auto is hard‑coded in run_in_container.sh.
~/dev/codex/codex-cli/scripts/run_in_container.sh --work-dir ~/dev/ai/codex "Hello world" --approval-mode suggest --model o3
....
Firewall configuration complete
Verifying firewall rules...
Firewall verification passed - unable to reach https://example.com as expected
Firewall verification passed - able to reach https://api.openai.com as expected
╭──────────────────────────────────────────────────────────────╮
│ ● OpenAI Codex (research preview) v0.1.2504251709 │
╰──────────────────────────────────────────────────────────────╯
╭──────────────────────────────────────────────────────────────╮
│ localhost session: <<Session_id>> │
│ ↳ workdir: ~/dev/ai/codex │
│ ↳ model: o3 │
│ ↳ provider: openai │
│ ↳ approval: full-auto │
╰──────────────────────────────────────────────────────────────╯
Additional information
Proposed fix:
- Delete the literal <code>--full-auto</code>; leave the quoting loop untouched.
Unclear:
- Why was full auto hard coded does run in container need to be always in full-auto mode for any reason I am missing ?
PR incoming. ;)
FYI: if this is merged fast I would like to make a second PR on the run_in_container.sh which shall enhance the container to also mount the ~/.codex/ into the container so user configs, instructions and sessions are reusable in the container env.
This issue has 1 comment on GitHub. Read the full discussion on GitHub ↗