document_id Canonical MCP Recheck

Verdict

DOCUMENT_ID_CANONICAL_MCP_ACCEPTED

canonical_document_id() rejects dot segments, empty segments, repeated/leading/trailing separators, backslash, URL-encoded aliases, non-ASCII/homoglyph paths, scope mismatch, and case variation, and requires byte-for-byte equality to the supplied MCP ID.

Acceptance is for this identity function and contract. Production integration remains blocked until the complete runnable sealing pipeline actually calls it for every consumed document.