fix: dag review node rejects matching fingerprints - identical hex reported as mismatch #410

Why

A review-decision node failed its result contract with review result fingerprint 2d660335439ae628 does not match current implementation 2d660335439ae628 - both sides are the IDENTICAL 16-hex string, yet the comparison rejected it. This killed workflow dag_fe5fdfe0f346GIs9LrGhw6UQEP (issue #404 continuation) at the final review gate and forced a parent-side manual review instead.

Evidence

• Wake message: Review result contract failed: review result fingerprint 2d660335439ae628 does not match current implementation 2d660335439ae628 (first 16 hex of sha256 of git diff HEAD~1; commit 40d510485 on fix/404-headless-init-does, 2 files, +198/-7)
• The verify node had independently recomputed and confirmed the same fingerprint
• The implementation was subsequently accepted by parent review and shipped as PR fix: headless init does not stamp project time_initialized so memory stays inert #405 / release v1.0.30 - the fingerprint values were correct

Suspects

String identity looks equal in the message; suspect a type mismatch (string vs wrapped object), whitespace/prefix normalization, or comparing against a stale snapshot field instead of the echoed one in the review-result contract check.

Acceptance

• Reproduce with a review node echoing the exact fingerprint string
• Fix the comparison to accept byte-identical fingerprints
• Regression test pinning the equal-fingerprint accept path