fix(schemas): add missing 'executive_synthesis_skipped' to Event.type enum#1
Open
MerouaneBen wants to merge 1 commit into
Open
fix(schemas): add missing 'executive_synthesis_skipped' to Event.type enum#1MerouaneBen wants to merge 1 commit into
MerouaneBen wants to merge 1 commit into
Conversation
… enum The reporter's executive-synthesis fallback emits Event(type='executive_synthesis_skipped', ...) when the exec-summary LLM call raises, but that value is absent from the Event.type Literal in schemas.py. Pydantic then raises ValidationError, crashing the entire run at the reporter stage — after the (expensive) jury + consensus work is already done. The graceful fallback is itself the failure path. Adding the enum value lets the run degrade to the deterministic summary as intended. Found while evaluating a fleet of internal agents; reproduced reliably when the harness LLM returns malformed JSON / times out on the exec-synthesis call.
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary
When the Reporter's executive-synthesis LLM call raises, the graceful-fallback path emits
Event(type="executive_synthesis_skipped", …)— but that string isn't in theEvent.typeLiteralenum (schemas.py). Pydantic then raisesValidationError, which crashes the entire run at the reporter stage, after the (expensive) jury + consensus work is already done. The safety net is itself the failure.Where
proofagent_harness/agents/reporter.py→_generate_executive_synthesis()except Exception:block calls_emit(state, Event(type="executive_synthesis_skipped", …)). TheEvent.typeLiteralinschemas.pydoes not list that value.Repro
Any
Harness().evaluate(...)where the harness LLM fails/refuses/times out on the exec-synthesis call. We hit it reliably with a--metricssubset + a harness model that occasionally returns malformed JSON. Instead of degrading to the deterministic summary, the process dies with:Fix
Add
"executive_synthesis_skipped"to theEvent.typeLiteral. Verified locally — the run then completes on the deterministic fallback as intended.Context
Found while running a serious hands-on evaluation of a fleet of internal agents. Genuinely useful harness — thanks for open-sourcing it.