Describe the feature or problem you'd like to solve
When running long-running commands (builds, tests, lints), the CLI provides no visibility into progress. Users see nothing until the command completes or times out, then receive a large dump of output. This makes it unclear whether the build is progressing, stuck, or failing early. Currently, streaming requires manual read_powershell calls which consume LLM turns and add overhead.
Proposed solution
Add automatic output streaming/tailing for long-running powershell commands.
Options could include:
- A stream: true parameter that periodically displays output to the user without LLM round-trips
- Automatic tail (e.g., last 20 lines every 10 seconds) shown in the UI while the command runs
- A progress indicator showing the command is still active
Benefits: Users gain confidence builds are progressing, can spot errors early, and get a more responsive experience. Reduces anxiety during multi-minute operations.
Example prompts or workflows
Example prompts or workflows (3-5):
- "Build chrome" — user sees incremental compile progress instead of waiting 5+ minutes blind
- "Run browser_tests" — user watches test counts increment (e.g., "[1500/3000] tests passed")
- "Run the full test suite" — early failures surface immediately rather than after 10 minutes
- "Start the dev server and verify it's working" — user sees server startup logs in real-time
Additional context
No response
Describe the feature or problem you'd like to solve
When running long-running commands (builds, tests, lints), the CLI provides no visibility into progress. Users see nothing until the command completes or times out, then receive a large dump of output. This makes it unclear whether the build is progressing, stuck, or failing early. Currently, streaming requires manual read_powershell calls which consume LLM turns and add overhead.
Proposed solution
Add automatic output streaming/tailing for long-running powershell commands.
Options could include:
Benefits: Users gain confidence builds are progressing, can spot errors early, and get a more responsive experience. Reduces anxiety during multi-minute operations.
Example prompts or workflows
Example prompts or workflows (3-5):
Additional context
No response