anomalyco/opencode
1
0
Fork 0
mirror of https://github.com/anomalyco/opencode.git synced 2026-07-22 10:15:31 -04:00
Code Issues 5.9k Packages Projects Releases 19 Wiki Activity

Custom commands don't appear in message history #2093

New Issue
Closed
opened 2026-02-16 17:34:07 -05:00 by yindo · 8 comments
yindo commented 2026-02-16 17:34:07 -05:00
Owner
Copy Link
Copy Source

Originally created by @fpv-dev on GitHub (Oct 14, 2025).

Originally assigned to: @rekram1-node on GitHub.

I created /subtask command to force opencode run my request as a subtask to not pollute main thread context. The problem is that my request disappear and I can't lookup it using arrow up key. Maybe I'm doing something wrong or misuse :) Another minor UX issue that there is no any notification or message that subtask is completed

---
description: run request as a subtask
subtask: true
---

$ARGUMENTS
Image
Originally created by @fpv-dev on GitHub (Oct 14, 2025). Originally assigned to: @rekram1-node on GitHub. I created `/subtask` command to force opencode run my request as a subtask to not pollute main thread context. The problem is that my request disappear and I can't lookup it using arrow up key. Maybe I'm doing something wrong or misuse :) Another minor UX issue that there is no any notification or message that subtask is completed ```md --- description: run request as a subtask subtask: true --- $ARGUMENTS ``` <img width="446" height="333" alt="Image" src="https://github.com/user-attachments/assets/b2ae1845-ace9-40f6-a1a7-ad6da2dc6d93" />
yindo closed this issue 2026-02-16 17:34:07 -05:00
yindo commented 2026-02-16 17:34:07 -05:00
Author
Owner
Copy Link
Copy Source

@github-actions[bot] commented on GitHub (Oct 14, 2025):

This issue might be a duplicate of existing issues. Please check:

  • #2232: Preserve Prompt on Ctrl+C - discusses saving prompts in history when cancelled
  • #1637: Linked files are broken when using history - describes issues with message restoration after restart
  • #1266: Bug: message history restoration corrupts text with attachments - details corruption of messages with attachments when using up arrow navigation
  • #314: history_previous and history_next keybinds not working - mentions up/down arrow keys not cycling through message history (though this was apparently fixed)

These issues all relate to message history functionality and restoration. The core issue you're experiencing (custom commands not appearing in history accessible via arrow up) seems closely related to the broader message history system that has had various bugs reported.

Feel free to ignore if none of these address your specific case with custom subtask commands.

@github-actions[bot] commented on GitHub (Oct 14, 2025): This issue might be a duplicate of existing issues. Please check: - #2232: Preserve Prompt on Ctrl+C - discusses saving prompts in history when cancelled - #1637: Linked files are broken when using history - describes issues with message restoration after restart - #1266: Bug: message history restoration corrupts text with attachments - details corruption of messages with attachments when using up arrow navigation - #314: history_previous and history_next keybinds not working - mentions up/down arrow keys not cycling through message history (though this was apparently fixed) These issues all relate to message history functionality and restoration. The core issue you're experiencing (custom commands not appearing in history accessible via arrow up) seems closely related to the broader message history system that has had various bugs reported. Feel free to ignore if none of these address your specific case with custom subtask commands.
yindo commented 2026-02-16 17:34:07 -05:00
Author
Owner
Copy Link
Copy Source

@rekram1-node commented on GitHub (Oct 14, 2025):

Brief explanation (caveat: we need to improve it haha)

yeah we need to show it better, the ux here is weird but basically what happens is we have a message telling the main agent that a task was fired off the main agent sees a truncated prompt send to subagent and then the result of the task run (the last text output of subagent)

You can see if it is still running my looking for working…. indicator it will disappear once task is completed

After the task completes the control goes back to you, you just succeeded in firing off a task now you can continue conversation however youd like

TLDR: we need to make things more clear, UX here isnt great

@rekram1-node commented on GitHub (Oct 14, 2025): Brief explanation (caveat: we need to improve it haha) yeah we need to show it better, the ux here is weird but basically what happens is we have a message telling the main agent that a task was fired off the main agent sees a truncated prompt send to subagent and then the result of the task run (the last text output of subagent) You can see if it is still running my looking for working…. indicator it will disappear once task is completed After the task completes the control goes back to you, you just succeeded in firing off a task now you can continue conversation however youd like TLDR: we need to make things more clear, UX here isnt great
yindo commented 2026-02-16 17:34:07 -05:00
Author
Owner
Copy Link
Copy Source

@rekram1-node commented on GitHub (Oct 14, 2025):

just curious, if you have any opinion here, what would you expect to happen?

I wanna make the ux better so figured since this is pretty confusing for users and you ran into it, you may have an opinion

@rekram1-node commented on GitHub (Oct 14, 2025): just curious, if you have any opinion here, what would you expect to happen? I wanna make the ux better so figured since this is pretty confusing for users and you ran into it, you may have an opinion
yindo commented 2026-02-16 17:34:07 -05:00
Author
Owner
Copy Link
Copy Source

@ferrao commented on GitHub (Oct 14, 2025):

Let me add my use case as well @rekram1-node , I have a command that fires a subagent that I want to have an interactive conversation with, but whatever the subagent asks me is not visible in the UI, unless I explicitly navigate to the subagent with CTRL+ LEFT.

Would it not make more sent to automatically switch to the session of whatever agent/subagent is working automatically and then let the user override it with the CTLR navigation instead?

@ferrao commented on GitHub (Oct 14, 2025): Let me add my use case as well @rekram1-node , I have a command that fires a subagent that I want to have an interactive conversation with, but whatever the subagent asks me is not visible in the UI, unless I explicitly navigate to the subagent with CTRL+ LEFT. Would it not make more sent to automatically switch to the session of whatever agent/subagent is working automatically and then let the user override it with the CTLR navigation instead?
yindo commented 2026-02-16 17:34:08 -05:00
Author
Owner
Copy Link
Copy Source

@rekram1-node commented on GitHub (Oct 14, 2025):

@ferrao I'm definitely of the opinion you should have some option of conversing with subagents, rn they are very limited as you have noticed.

I think there are a lot of easy wins we need to knock out to make the subagents more useful

@rekram1-node commented on GitHub (Oct 14, 2025): @ferrao I'm definitely of the opinion you should have some option of conversing with subagents, rn they are very limited as you have noticed. I think there are a lot of easy wins we need to knock out to make the subagents more useful
yindo commented 2026-02-16 17:34:08 -05:00
Author
Owner
Copy Link
Copy Source

@ferrao commented on GitHub (Oct 14, 2025):

they are very limited as you have noticed.

I wanted to kick off an interactive conversation with a research sub-agent that I created, define some completion checklists and when everything was ready, it should write out the results of said research into a set of structured files, that would then be read by other agents to continue the work.

But I am indeed realizing that I can not achieve this @rekram1-node

@ferrao commented on GitHub (Oct 14, 2025): > they are very limited as you have noticed. I wanted to kick off an interactive conversation with a research sub-agent that I created, define some completion checklists and when everything was ready, it should write out the results of said research into a set of structured files, that would then be read by other agents to continue the work. But I am indeed realizing that I can not achieve this @rekram1-node
yindo commented 2026-02-16 17:34:08 -05:00
Author
Owner
Copy Link
Copy Source

@deltasquare4 commented on GitHub (Oct 28, 2025):

I have the same problem, but the use case is slightly different. I am using commands that do specific work as part of my workflow, which accept arguments. For instance /brainstorm Help me figure out how to achieve X by integrating Y into the code trigger a conversation. Sometimes, this goes in the wrong direction because I have not provided enough context and I want to edit my last command. I believe this may happen with or without subagents.

This is especially painful when I have referenced a bunch of files in my command arguments. Copy-pasting them does not bring back the reference because of #1266.

@deltasquare4 commented on GitHub (Oct 28, 2025): I have the same problem, but the use case is slightly different. I am using commands that do specific work as part of my workflow, which accept arguments. For instance `/brainstorm Help me figure out how to achieve X by integrating Y into the code` trigger a conversation. Sometimes, this goes in the wrong direction because I have not provided enough context and I want to edit my last command. I believe this may happen with or without subagents. This is especially painful when I have referenced a bunch of files in my command arguments. Copy-pasting them does not bring back the reference because of #1266.
yindo commented 2026-02-16 17:34:08 -05:00
Author
Owner
Copy Link
Copy Source

@rekram1-node commented on GitHub (Nov 17, 2025):

should be fixed on 1.0.69

@rekram1-node commented on GitHub (Nov 17, 2025): should be fixed on 1.0.69
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
Clear labels
acp
bad
bug
discussion
docs
documentation
enhancement
extension
good first issue
growth
help-wanted
mcp
model-problem
needs:compliance
nix
opentui
perf
pull-request
Mirrored from GitHub Pull Request
 tui
web
windows
zen
No Label
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
yindo
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: anomalyco/opencode#2093
Delete Branch "%!s()"

Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?