Fix native codex subagent model labels - #401
Merged
Merged
Conversation
(cherry picked from commit 857155c)
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Codex subagents could show the main agent’s model and reasoning effort instead of their own. This happened when the information sent when starting a subagent did not include those details.
Tcode now asks for each subagent’s own model and reasoning effort when it starts working. Labels stay hidden until that information arrives. If it arrives after the subagent finishes or is interrupted, the labels update without changing its status or summary.
The updated test covers working, completed, and interrupted subagents. It checks that Tcode asks for the right thread, does not borrow the main agent’s values, and keeps the subagent’s status when a reply arrives late.
Implemented with GPT-6 Astra in Tcode