Repository navigation
feat(fusion): let judge use its own knowledge and override the panel - #6804
Merged
diegosouzapw merged 1 commit intoJul 10, 2026
Merged
diegosouzapw merged 1 commit into
diegosouzapw merged 1 commit into
Conversation
Loading
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Problem
buildJudgePrompt(open-sse/services/fusion.ts) tells the judge to write an answer "grounded in that analysis" — implicitly capping output at the panel's union. When every panel member misses something, or the consensus is collectively wrong/outdated, the judge is nudged to stay within their combined answer instead of correcting it.Change
Reword the judge prompt so the judge acts as a full participant: use its own reasoning/knowledge, override consensus when wrong, add what the panel missed, side with a correct lone source — with an explicit honesty guard against asserting low-confidence facts. Panel responses are still embedded and anonymized exactly as before.
Test
Adds
tests/unit/fusion-judge-own-intelligence.test.ts— asserts panel answers stay embedded, the old "grounded in that analysis" ceiling is gone, and the own-reasoning/override/honesty clauses are present. Verified failing on the old prompt, passing on the new (Hard Rule #18).Thanks for the great work on OmniRoute!