Feature Request: Disable attempt_completion tool #406
deathtothenig
started this conversation in
Feature Requests
Replies: 2 comments 5 replies
-
This is a must-have feature. Once you try it, you will agree. |
Beta Was this translation helpful? Give feedback.
5 replies
-
Yes it does, I don’t use this now. When I did I did this only for
orchestrator mode.
…On Thu, 29 May 2025 at 5:39 AM, Danila Sentyabov ***@***.***> wrote:
this worked for me /newtask do not call attempt_completion tool without
asking me first
This will break automated "boomerang tasks," though.
—
Reply to this email directly, view it on GitHub
<#406 (reply in thread)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/BPOUQW2PLW6KJKVA3QFV34D3AZF3RAVCNFSM6AAAAABVM3PQTOVHI2DSMVQWIX3LMV43URDJONRXK43TNFXW4Q3PNVWWK3TUHMYTGMZQGM3TONQ>
.
You are receiving this because you commented.Message ID:
***@***.***>
|
Beta Was this translation helpful? Give feedback.
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
-
This is the first thing I remove when I update Roo Cline. I hate it. I was using claude.ai chat before moving to this extension and you could have an ongoing conversation where Claude retains context for related tasks without him constantly trying to end the conversation by marking it as complete.
With the attempt_completion tool, Cline is constantly marking tasks as completed when they aren't. There are either problems that need to be fixed or the task has other things that weren't finished.
All I do is remove the code for the attempt_completion tool, but you would want to add an option to disable it in the settings. It's good for all modes, because even in the ask and architecture modes, which are amazing and solved one of the other problems I had, which was Cline running off and coding something I didn't ask for when I just wanted to discuss how to do something, I believe Cline will mark a task as complete after he has answered your original question.
The experience is 10x when you can just keep an ongoing chat without ever having Cline try to end it with the green wall of text. I also modified the prompt so that instead of trying to use the attempt_completion tool Cline is instructed to ask if everything works or if the user has any questions about what was done and that's what Cline does now when he thinks he has completed a task or answered my question.
Beta Was this translation helpful? Give feedback.
All reactions