fix(loop): detect repeats before committing assistant turn - #62
Merged
Conversation
When the loop detector fired after pushing the assistant message, the persisted history ended on a tool_use with no matching tool_result. On session resume the next request carried that dangling pair and was rejected by the model API. Move repeat detection ahead of the history.push so a detected repeat returns history without committing the turn. Skip the check on truncated calls, which are already steered to splitting via splitWriteHint; a repeat on the next turn there means 'try again smaller', not 'loop detected'. Adds a vitest suite pinning the tool_use -> tool_result block-ordering invariant and the various guard paths (truncation, unknown tool, denied permission, invalid input, leaked text call, repeated tail).
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
When the loop detector fired after pushing the assistant message, the persisted history ended on a tool_use with no matching tool_result. On session resume the next request carried that dangling pair and was rejected by the model API.
Move repeat detection ahead of the history.push so a detected repeat returns history without committing the turn. Skip the check on truncated calls, which are already steered to splitting via splitWriteHint; a repeat on the next turn there means 'try again smaller', not 'loop detected'.
Adds a vitest suite pinning the tool_use -> tool_result block-ordering invariant and the various guard paths (truncation, unknown tool, denied permission, invalid input, leaked text call, repeated tail).