fix(retrain): soft sanity gate for the retrain workflow - #17
Merged
Conversation
The retrain workflow ran the full pytest suite (incl. the frozen golden snapshot) to verify fitted weights. But any weight change alters the detector's exact output, so the snapshot trips by design — meaning the workflow could never propose an improvement without a red run. Replace it with a lightweight sanity check (clear-AI high, clear-human low, adequate gap) marked continue-on-error; the PR only opens when it passes. The proposed PR's own CI (and a note in its body) is where the golden snapshot gets refreshed by a human. 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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
The retrain workflow failed because its verify step ran the frozen golden snapshot, which trips on any weight change by design — so it could never propose an improvement. Replaced with a soft sanity check (AI high / human low / adequate gap,
continue-on-error); the PR opens only when it passes, and its own CI is where a human refreshes the golden hash.🤖 Generated with Claude Code