Skip to content

[FEATURE] The ability to queue up messages.via asking the user for further info on user input. #34835

Description

@horos22reg

Preflight Checklist

  • I have searched existing requests and this feature hasn't been requested yet
  • This is a single feature request (not multiple features)

Problem Statement

This is close to a dup (not truly, as as you'll see below I have a twist on the following request- #535 - where it is stated in the ticket that there is a fix but it seems to have ended up in the great bit bucket in the sky.

Basically the goal is to have a feature similar to cursor's - namely to be able to queue up messages for claude to work on rather than replace the current workflow. As.it stands I am trying to push for claude code in my work but when I get messages like 'hey - my workflow was interrupted what's going on?.. and 'I just wasted HOW many hours because claude did WHAT when I told it to do this additional task?'. this does not make my job easy.

Basically the idea is this. When you type an additional instruction for claude while it is working on something currently, instead of going ahead and killing its current thought process as currently exists, it asks you a clarifying question:

Do you want to (s)top the current workflow or (q)ueue up a new task?

And it replaces or queues as necessary. That way it gets rid of the ambiguity in the prompt. As it stands, it is not clear here whether you are actually wanting to replace the current task or add to it. Adding this prompt gets rid of the ambiguity adds and extra feature to claude code, and makes the user's intent clear.

@8enmann - you said that you had a fix for this in issue 535. I must be missing something here, but I can't imagine that this would be too much effort and would love to provide a patch.

Proposed Solution

Basically the idea is this. When you type an additional instruction for claude while it is working on something currently, instead of going ahead and killing its current thought process as currently exists, it asks you a clarifying question:

Do you want to (s)top the current workflow or (q)ueue up a new task?

And it replaces or queues as necessary. That way it gets rid of the ambiguity in the prompt. As it stands, it is not clear here whether you are actually wanting to replace the current task or add to it. Adding this prompt gets rid of the ambiguity adds and extra feature to claude code, and makes the user's intent clear.

@8enmann - you said that you had a fix for this in issue 535. I must be missing something here, but I can't imagine that this would be too much effort and would love to provide a patch.

Alternative Solutions

You can emulate this by telling claude to write to a done file when completed in your request in one window, then spawning up another claude instance and adding something like:

first, wait until this done file exists. Then, do the following.

But this is inadequate. the second claude instance doesn't contain the first claude instance's context for one. Second its iffy because it requires the user to think ahead of time to put this statement in, and it isn't bulletproof because it isn't a hardcoded instruction.

Priority

High - Significant impact on productivity

Feature Category

CLI commands and flags

Use Case Example

Example scenario:

I'm telling claude to monitor a production run for anomalous occurances.
I get about an hour in, and remember that I forgot to tell claude to do something afterwards (clean up the filesystem, etc)

I'm now stuck either with restarting the prompt from scratch, or somehow jerry-rigging a method of spawning claude to do the extra work by watching for the end of a process in the process table, or what not.

Additional Context

Please fix this promptly. Its been sitting here for about a year now, with several people commenting on and saying how great this feature would be to have. Again, maybe I'm missing something but I'm having a hard time thinking of how this could be a difficult fix to make.

Metadata

Metadata

Assignees

No one assigned

    Labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions