codex-worker

Codex Worker

Run Codex CLI as the implementation worker for coding tasks.

How to Use

Do not pass free-form user text to the shell unquoted. The task may contain spaces or shell metacharacters such as (, ), ;, #, *, or [.

If the user only supplied a task or instructions, execute:

"${CLAUDE_PLUGIN_ROOT}/scripts/codex-worker.sh" "$ARGUMENTS"

If the user supplied flags such as --worker-model, --worker-timeout, or --workdir, reconstruct the command so those flags remain separate shell arguments and the remaining free-form task is passed as one quoted final argument.

Example:

"${CLAUDE_PLUGIN_ROOT}/scripts/codex-worker.sh" --worker-model gpt-5.4:xhigh --workdir /repo/worktree "Implement the pending coding task and run relevant checks"

Never run this unsafe form:

"${CLAUDE_PLUGIN_ROOT}/scripts/codex-worker.sh" $ARGUMENTS

because the shell will re-parse the task text and can fail before codex-worker.sh starts.

Output

• The script outputs the worker response to stdout and status info to stderr
• Read the stdout output carefully and incorporate the worker result into your coordination flow
• If the script exits with a non-zero code, report the error to the user

Error Handling

Exit Code	Meaning
0	Success - worker response is in stdout
1	Validation error (missing codex, empty task, invalid flags, empty response)
124	Timeout - suggest using --worker-timeout with a larger value
Other	Codex process error - report the exit code and any stderr output

Notes

• Default worker model is gpt-5.4:xhigh
• The response is saved to .humanize/skill/<timestamp>/output.md for reference