Repository navigation
fix(mars): route verification and refinement using matching agent IDs - #346
Merged
codelion merged 1 commit intoOct 11, 2026
Conversation
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
MARSAgentuses integer IDs, butgenerate_solutionstores them as strings. The verifier compares these values directly, so agent 0 can verify its own answer, the temperature-diversity preference cannot find the author, and both improvement paths silently skip generated solutions because no original agent matches.Compare IDs as strings when selecting a verifier and locating the author for refinement. This preserves integer-ID compatibility and the existing fallback for aggregated solutions or a single agent.
The regressions use real generation, verification, and improvement methods with mocked model responses. They cover all three authors in sequential and parallel mode, checking the selected peer, feedback passed to the author, updated answer, and token accounting. All six pipeline cases fail before the fix.
Validation:
python -m pytest tests/test_mars_agent_routing.py tests/test_mars_parallel.py tests/test_mars_workspace_verification.py tests/test_mars_aggregation_sampling.py -qpassed 26 tests and 4 subtests. Ruff on the new tests andgit diff --checkpassed. No live model calls.