Feature request: Claude Code-like /monitoring for background terminals
Open 💬 2 comments Opened Apr 14, 2026 by wolverin0
💡 Likely answer: A maintainer (github-actions[bot], contributor)
responded on this thread — see the highlighted reply below.
Summary
Request a Claude Code-like /monitoring workflow for background terminals in Codex.
This includes:
- a dedicated monitored view for running background terminals
- readiness / warning / error pattern detection surfaced in the TUI
- a targeted stop action for a single background terminal by
process_id - a slash-command UX similar to Claude Code-style monitoring flows
Proposed UX
/monitorsto show monitored background terminals/monitor-stop <process_id>to stop one monitored terminal- inline TUI notifications when a monitored terminal emits ready / warning / error signals
Implementation branch
A working implementation was prepared here:
- compare URL: https://github.com/openai/codex/compare/main...wolverin0:feat/monitor-background-terminals?expand=1
- fork branch:
wolverin0:feat/monitor-background-terminals - commit:
535e61f
Notes
PR creation is currently blocked by permissions for openai/codex, so this issue is being opened first to track the feature request and proposed implementation.
2 Comments
Potential duplicates detected. Please review them and close your issue if it is a duplicate.
Powered by Codex Action
A /monitoring style view gets much more useful if it shows why a process is still worth waiting on, not just that it is alive. The fields I keep wanting are started_at, last_output_at, last_ready_signal_at, ^[rror_count, and allback_budget_remaining, because the hard part is deciding when the background terminal has become dead weight instead of productive background work. That is the same seam we have been hitting in MartinLoop around long-running coding-agent loops.