Feature Request - Mark task as completed (needs scoping)

[James-Cherished]

App Version

3.10.4

API Provider

OpenRouter

Model Used

Any

Actual vs. Expected Behavior

Hi!

I like to see the green text that shows the task successfully solved the issue/request when managing my tasks.
Sometimes the issue/request is solved but Roo is still waiting for your input/confirmation. I state the task is complete, which allows me for future reference (debugging, pinpointing errors, etc.. for example). But this might cost me a few API calls, as Roo tries new things, attempts commands, asks questions, or further monologues before marking the task complete.

Manually closing a task as complete would allow for both future referencing and starting a new task for the next thing right away, saving on cost.

Thanks a lot!

Detailed Steps to Reproduce

1. Create and run a task
2. Stop mid-task / the task is waiting user action
3. You wish you could mark the task as complete

Relevant API Request Output

Additional Context

No response

[James-Cherished]

in the same spirit, it might also be useful to mark them as failed, with red text, actually. Highlights roadblocks / dead ends

[hannesrudolph]

Is this in reference to using boomerang mode?

[James-Cherished]

Hi! No, not specifically. Do you need clarification?

[hannesrudolph]

@James-Cherished what models are you seeing this on? This seems like a model limitation in terms of understanding the current context of the chat. The LLM is the brains and Roo is the orchestrator.

[James-Cherished]

Thanks for the answer!

Maybe it was too late when I wrote this 😅 💤 To be more synthetic :

• Feature request : Manually mark tasks as complete or failed, while running mid-task
• Why : saves API calls, useful for personal task management (cut a task short, mark it as complete or dead end)

Anyways, not the most important thing at all, just sharing my thoughts as I use it, maybe this one's too personal of a need.

[hannesrudolph]

Thanks for the answer!

Maybe it was too late when I wrote this 😅 To be more synthetic :

• Feature request : Manually mark tasks as complete or failed, while running mid-task
• Why : saves API calls, useful for personal task management (cut a task short, mark it as complete or dead end)

Anyways, not the most important thing at all, just sharing my thoughts as I use it, maybe this one's too personal of a need.

I like where you're going with this. Roo does not have a built in task management framework or modes yet but it is something we are looking to implement. Your input is valuable. Thank you.