Skip to content

classifier-fingerprint: looks_like_ai drift from PR #3537 not acknowledged #3540

Description

@Sinity

Summary

`devtools lab policy classifier-fingerprints` reports an undeclared fingerprint drift for `polylogue/sources/parsers/claude/ai_parser.py:looks_like_ai`, discovered while verifying an unrelated PR (polylogue-3uw, capture-completeness coverage) whose pre-push quick-verify gate runs this check repo-wide regardless of the files that PR touches.

Evidence

Ask

Determine whether #3537's `looks_like_ai` tightening needs an `INDEX_SCHEMA_VERSION` bump with a declared `SEMANTIC_REPARSE` delta (already-indexed Claude-AI-adjacent rows may be classified differently under the tightened rule), or whether it's safe as a retroactive `acknowledged_safe` entry (matching the existing `polylogue-zqph` precedent for the same class of gap from PR #3428). Until resolved, every push in the repo fails `devtools verify --quick``'s `lab policy classifier-fingerprints` step, since it is repo-wide and un-scoped to the changed files of the pushing branch.

Discovered while working polylogue-3uw; unrelated to that bead's scope.

Activity

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions