fix(pipeline): honest failure classification (infra vs validation) + CLI validate no longer false-PASS#110
Merged
Merged
Conversation
…top CLI validate false-PASS
The orchestrator's RunAgentResult folded registration failures
(contract_missing, agent_missing) into the same shape as genuine gate
rejections, so a consumer could not tell "OTAIP wasn't wired to validate
this" from "the input was rejected" — bad setup could read as a model
regression.
- Add FailureClass ('infra' | 'execution' | 'validation') and
classifyFailure(); attach failureClass to every RunAgentResult failure.
Re-export from the pipeline-validator barrel and @otaip/core.
- CLI `validate` previously always printed Overall: PASS (every gate
hardcoded passed:true). It now reports gates it never ran as SKIPPED,
marks the verdict UNVALIDATED (infra), and exits non-zero — a
non-validation can no longer read as a pass.
- Tests: exhaustive classifyFailure table; orchestrator infra/execution
cases; CLI test asserting validate never reports PASS.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Why
Field feedback flagged the worst property of using OTAIP as an oracle: infra errors look like model failures. Two real, OTAIP-owned weaknesses caused it.
What
FailureClass = 'infra' | 'execution' | 'validation'+classifyFailure(), and afailureClassfield on everyRunAgentResultfailure.contract_missing/agent_missing→infra;agent_error→execution; gate rejections →validation. Re-exported from@otaip/core. Additive — no change to the success path.validate: previously hardcoded every gate topassed: trueand always printedOverall: PASS. Now reports unrun gates asSKIPPED, verdictUNVALIDATED (infra), and exits non-zero. A non-validation can no longer read as a pass.Tests
Exhaustive
classifyFailuretable; orchestrator infra/execution/validation cases; CLI test assertingvalidatenever reports PASS for an uncontracted agent. Full suite green.🤖 Generated with Claude Code