[Bug]: Needs clarification - give more visibility into whether setup.sh was successful

[openhands-agent]

Describe the bug and reproduction steps
When setup.sh runs, it may or may not succeed. If it fails, we should show a message to the user. Would be nice to see the logs etc

Potential solutions Maybe we can print the command output to a file and show that in the first message?

Logs, Errors, Screenshots, and Additional Context
<N/A>

Issue Created By: Robert Brennan on Slack
Link to Slack Thread: https://openhands-ai.slack.com/archives/C08HMC4S5A6/p1744309826954479?thread_ts=1744309826.954479&cid=C08HMC4S5A6

[github-actions]

This issue is stale because it has been open for 30 days with no activity. Remove stale label or comment or this will be closed in 7 days.

[jpshack-at-palomar]

Important in the context of all-hands cloud.

It would be great to see output of the setup.sh via a logging interface or via the integrated shell. The shell would be most natural but I imagine you had reasons for not doing it there--out of curiosity why not run setup in the context of the shell visible to users in the UI?

BTW I have seen odd behavior with respect to setup.sh that appears under some conditions to prevent the agent from properly initializing at all so that it either never comes up at all or comes up but then when it starts to execute commands complains that there is some process already running that it can't kill, etc. Without logs I can't see what is going on. And it seems that if it ran after startup in the normal terminal, we would be less likely to get into an unrecoverable state.

[neubig]

@OpenHands let's take a stab at making this better. when running openhands with the UI, it would be ideal to actually display the command that is being run chmod +x .openhands/setup.sh && source .openhands/setup.sh in the UI, and also show the result of this command in the terminal. propose any changes that would need to be made to the codebase to make this possible.

[openhands-ai]

I'm on it! neubig can track my progress at all-hands.dev

[openhands-ai]

Summary

I have successfully implemented the feature to display setup.sh command execution and output in the OpenHands UI, addressing GitHub issue #7797.

What Was Implemented

🎯 Core Feature

• Command Visibility: The exact command chmod +x .openhands/setup.sh && source .openhands/setup.sh is now displayed in the UI as a collapsible action block
• Output Visibility: The complete output of the setup script execution is shown in the UI as an observation
• Status Feedback: Clear success/failure messages are provided to users

🔧 Technical Changes

Backend (openhands/runtime/base.py):

• Modified maybe_run_setup_script() to add setup command events to the event stream
• Added CmdRunAction with descriptive thought: "Running setup script to configure the workspace environment."
• Added CmdOutputObservation to show command output in the UI
• Used EventSource.ENVIRONMENT for proper event categorization
• Enhanced status callbacks with translated success/failure messages

Frontend:

• Added multilingual translation support for setup script status messages in 16 languages
• Updated type definitions for new translation keys

Testing:

• Added comprehensive test coverage for both successful and failed setup script executions
• Verified proper event stream integration and UI visibility

🌟 Benefits

• Transparency: Users can see exactly what setup commands are being executed
• Debugging: Setup script output is visible for troubleshooting failed setups
• User Experience: Consistent with how other commands are displayed in OpenHands
• Multilingual: Status messages available in 16 languages

📋 Quality Assurance

• ✅ All pre-commit hooks pass (formatting, linting, type checking)
• ✅ Comprehensive test coverage added
• ✅ Frontend translation completeness verified
• ✅ Changes are minimal and focused on the specific requirement

Pull Request Created

PR #9318: #9318

The implementation leverages OpenHands' existing event stream infrastructure, ensuring seamless integration with the current UI without requiring major architectural changes. Users will now have full visibility into setup script execution, making debugging and troubleshooting much easier.