mathiasandClaude Sonnet 5
442557ca31
Round 7's review found an 8th bypass in the same "value hidden from the extractor" family as round 6, inside the fold logic round 7's own fix had just added (a blank line interrupting a folded continuation) — the second distinct gap found on the first alternate shape tried against that fix. Decision (Mathias, via the running sweep session): ship this PR's rounds 1-7 as v1 rather than continue incremental patches past the point of diminishing returns. Every bypass found across all 7 rounds requires the caller to already hold brain_edit write access and to deliberately construct a specific malformed-YAML shape with knowledge of this exact gap — narrow and low-likelihood next to another patch cycle. The real fix (one shared, tested, encoding-normalized frontmatter parser, replacing the three independent hand-rolled implementations this bug class keeps surfacing in) is tracked separately as hyperguild#138. Document the boundary directly on checkClaimEnvelope so the next person reading this function knows what's covered and what's a deliberate, recorded trade-off rather than an unnoticed gap. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>