/goal when blocked doesn't stop but keeps looping consuming tokens and API credits
What version of Codex CLI is running?
codex-cli 0.130.0
What subscription do you have?
PRO
Which model were you using?
gpt-5.5
What platform is your computer?
Windows 11
What terminal emulator and version are you using (if applicable)?
gitbash
Codex doctor report
not available
What issue are you seeing?
in /goal mode, when the goal is blocked requiring "operator" decision/input, it loops repeatedly. If the operator is away or run overnight, this loop continues indefinitely, compacting the context, still continuing, consuming API credits for ever multiple times.
What steps can reproduce the bug?
Uploaded thread: 019e3318-2b97-79d3-bc3d-1bca34d7ed3a. Here is the codex log output loop many such lines.
Blocked on A authorization. I need explicit approval to cancel stale production orders 7, 7, and 7 before continuing.
Blocked pending explicit A production mutation approval. Required approval: cancel stale A orders 7, 7, and 7*.
Blocked on A authorization. The next action is cancelling stale production orders 7, 7, and 7; I need explicit approval before running it.
Blocked pending explicit operator approval. Required A mutation: cancel stale production orders 7, 7, and 7.
Blocked on A. I need explicit approval before mutating production state to cancel orders 7, 7, and 7**.
What is the expected behavior?
When blocked, /goal should pause , and wait for user to explicitly /goal resume.
Additional information
_No response_
This issue has 2 comments on GitHub. Read the full discussion on GitHub ↗