feat(run): propagate control failureClass onto canonical RunRecord (agent-eval 0.54.0)#71
Merged
Merged
Conversation
…gent-eval 0.54.0) agent-runtime built RunRecords but never set the structured failure tag, starving the substrate's cross-agent failure aggregation. The control layer already classifies failures (ControlRunResult.failureClass) — this promotes that onto the canonical RunRecord.failureClass key (agent-eval 0.54.0) at the projection boundary, so failures aggregate across every agent in ONE taxonomy vocabulary. - bump @tangle-network/agent-eval ^0.52.0 -> ^0.54.0 - applyRunRecordDefaults(): stamps scenarioId + failureClass without overriding anything the adapter set; promotes ONLY real FAILURE_CLASSES taxonomy tags (a non-taxonomy string like "forge_build_unsatisfied" stays unclassified, not polluting the canonical key) - extracted as a pure exported helper + 4 unit tests (promote / don't-override / reject-non-taxonomy / no-failure) typecheck clean, 374/374.
tangletools
pushed a commit
that referenced
this pull request
May 28, 2026
…gents-of-all-shapes showcase (#72)
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary
@tangle-network/agent-eval^0.52.0→^0.54.0and consumes the new canonical cross-agentRunRecord.failureClass.ControlRunResult.failureClass) but it never reached the canonical record — so substrate failure aggregation was starved on the runtime side. This promotes it at the projection boundary.applyRunRecordDefaults()stampsscenarioId+failureClasswithout overriding anything the adapter set, and promotes only realFAILURE_CLASSEStaxonomy tags (a non-taxonomy string likeforge_build_unsatisfiedstays unclassified rather than polluting the canonical key).Why
Closes the runtime half of the unified failure-classification architecture (agent-eval 0.54.0): every agent's failures now aggregate in ONE vocabulary, making "which failure dominates across the whole fleet" answerable, with per-agent detail preserved in
failureMode.Test plan
pnpm typecheckcleanpnpm test— 374/374 (4 newapplyRunRecordDefaultsunit tests: promote canonical / don't-override-adapter / reject-non-taxonomy / no-failure)pnpm buildclean