chore: remove verify script dead code - #4102
Closed
melmennaoui wants to merge 1 commit into
Closed
Conversation
Remove the verify field from EvalCriteria and all supporting code. The verify concept (running a shell script via docker exec on the eval container after the agent completes) cannot work without container lifecycle changes: the container uses --rm and is deleted before verify could run. The assertion system already covers verifiable cases — response text, tool calls, cost — via the captured event stream. Rather than ship dead code that creates user confusion (the schema accepts it, nothing runs it), remove it entirely. The schema's DisallowUnknownFields rejection ensures a future verify field can be added cleanly when a real need is proven. Removed: - EvalCriteria.Verify field - VerifyCheck type and EvalResultChecks.Verify field - cloneEvalResultChecks verify branch - verify.go and verify_test.go - verify test case in session_test.go
melmennaoui
force-pushed
the
chore/remove-verify-dead-code
branch
from
September 1, 2026 15:55
ca9c9e3 to
3805fc3
Compare
Contributor
Author
|
Folded into #4103 |
aheritier
added
kind/chore
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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
Remove the
verifyfield fromEvalCriteriaand all supporting code. This was shipped in #4088 and #4089 but is dead code — it cannot work without container lifecycle changes that aren't warranted.Why
The verify concept (run a shell script via
docker execon the eval container post-agent) requires the container to stay alive after the agent exits. The current pipeline uses--rm, so the container is deleted before verify could run.The assertion system already covers the verifiable cases:
contains,not_contains,equals,regex,starts_with,ends_withtool_calledcost_thresholdAll of these operate on the captured event stream — no live container needed.
Removed
EvalCriteria.VerifyfieldVerifyChecktype andEvalResultChecks.VerifyfieldcloneEvalResultChecksverify branchpkg/evaluation/verify.goandverify_test.gosession_test.goForward compatibility
EvalCriteria.UnmarshalJSONusesDisallowUnknownFields, so averifyfield in an eval JSON will produce a clear error. If a real need emerges later, adding the field back is a one-line change.