Repository navigation
[FR] Tweak plan then continue #2886
Description
After Claude proposes a plan, it gives the user two options: "would you like to proceed?: "yes" and "no, keep planning"
It would be helpful to add a third option "add comments, then proceed", which allows me to write a few small comments, then proceed in auto-accept mode.
At the moment, my options are:
- Press "yes", then immediately
escto interrupt, add my comment, then proceed, which is a bit annoying. - Press "no, keep planning", enter my comment, and wait for the entire plan to be re-generated, then press yes.
Both of those interactions add extra unnecessary tokens.
Activity
sid374 commentedon Jul 2, 2025
If you hit "no" and then give it your comments, it will regenerate the plan according to your comments and continue on. i.e even if you press no, Claude can still access the plan and take your comments in stride.
lycarter commentedon Jul 2, 2025
That makes sense, and that's my current workaround, but it's a few extra seconds of waiting and a few extra keystrokes.
Beyond the small time hit, my understanding is that when I press no, and it regenerates the plan, both plans need to be generated (more output tokens), and both plans wind up in context (at least, that's what it looks like visually, let me know if that's not actually what's happening).
For example, I ask it to perform a big refactor in a FE codebase. The plan is like...60 lines long, and one block of the plan is "update import order to be less confusing" or something along those lines. The tweak I want to give is "don't bother touching the import order, just run
prettierbefore committing and it'll auto-reorder". I would imagine that the nicely formatted plan plus my small additional comment are enough for it to proceed in auto-accept mode, without needing to regenerate a plan.sid374 commentedon Jul 2, 2025
Gotcha, if it's just a small tweak to the plan you could use realtime steering to adjust the plan after the model's started working on it. So this is what it would look like:
lycarter commentedon Jul 2, 2025
Will give that a shot, thanks!
RomainGehrig commentedon Jul 3, 2025
Just to add to the conversation, it would be nice to be able to pop the plan in $EDITOR, make some changes, save and close the file. Then Claude could look at it and go with it or ask clarifications by re-issuing the plan.
My use-case is more for longer plans around a full feature that just need some tweaks. If there are more than 2 things to change, I tend to forget what I wanted to change while typing the first few (and Claude is still pumping code in the background).
(If the whole plan needs tweaking, I just recreate a plan with different directions.)
Hyperrick commentedon Jul 5, 2025
It would be helpful if pressing "Esc" wouldn't delete the plan from the terminal output. Then I wouldn't need to take notes..
github-actions commentedon Aug 16, 2025
Found 3 possible duplicate issues:
Yes, and...,Yes, but...#2504This issue will be automatically closed as a duplicate in 3 days.
🤖 Generated with Claude Code
github-actions commentedon Aug 20, 2025
This issue has been automatically closed as a duplicate of #4039.
If this is incorrect, please re-open this issue or create a new one.
🤖 Generated with Claude Code
github-actions commentedon Aug 27, 2025
This issue has been automatically locked since it was closed and has not had any activity for 7 days. If you're experiencing a similar issue, please file a new issue and reference this one if it's relevant.