Adversarial appendix: prior work and the proposed bridge

25 September 2026. Independently read the new positioning memo and source ledger, then retrieved selected primary specifications rather than assuming the author’s 16-page inspection established every proposed mapping. This review verified the key distinctions below; it did not audit all implementations, recheck all 16 sources or demonstrate network interoperability.

Semantic crosswalk survives with explicit boundaries

• COAR Announce Endorsement 1.0.1 points to an endorsement resource and the endorsed resource separately. The memo correctly restricts this mapping to genuine endorsements. A local choice to use a calculation or an institution’s legal decision cannot silently become scholarly endorsement because a transport pattern exists.
• COAR Undo Offer acts on the earlier offer. The memo correctly refuses to treat it as automatic retraction, downstream amendment resolution or revocation of every use. Its instruction to report unsupported transport semantics is preferable to a lossy universal mapping.
• PROV-O already provides responsible agents, qualified associations, revisions and invalidation. The memo acknowledges these rather than inventing a provenance gap. PROV invalidation concerns an entity becoming unavailable for use; a disputed interpretation or one expired reliance does not automatically imply invalidation of every underlying evidence entity. Choosing the correct entity granularity remains an adapter design decision.
• RO-Crate profiles provide a genuine route for domain conventions. Its provenance guidance already records workflow changes, responsible human authorization and old/new files. The direct provenance request initially failed; following the official profile page’s link retrieved the relevant text. No present MCRP conformance follows, and the memo explicitly says no such adapter has been demonstrated.
• PCI’s FAQ and its version-controlled Registered Reports account substantiate existing editorial decisions, archived review records and version-aware workflows. The latter is an explanatory account, not a finding about this packet’s current eligibility or acceptance. The memo correctly treats these as prior capabilities, not empty spaces awaiting MCRP.

The proposed round-trip projection is an explicit future oracle, not an existing theorem. Even a successful semantic round trip does not transfer the originating institution’s authority to its recipient. Neither a persistent identifier nor contributor role establishes independent control or appointment.

Concrete repairs applied, coordinated with the author

1. Public reading path: the physics row referred readers to private dossier
   6. It now links to the public packet’s mathematical foundations and coupled allocation manuscripts. This makes the promised next step available in the release being assessed.
2. Human outcome framing: the biology row proposed improvement “at matched effort.” It now uses the same information and comparable declared support, reports actual total labor, and includes appropriately retained supported uses. This matches the prospective design and avoids rewarding blanket refusal.
3. Existing records are not automatically shareable: PCI’s FAQ explicitly distinguishes public recommended-review material from rejected reports that are not published. The invitation to bring an existing record and keep it available could otherwise encourage disclosure of confidential peer review. The memo now limits public examples to synthetic or permission-cleared records, preserves the permitted audience and instructs an adapter to expose or reject incompatible access restrictions. A person’s permission to read is not a permission to republish.

Only positioning prose changed; no code, tests, models, licenses or deployment instructions were modified. The paper’s narrower contribution claim remains: a particular cross-system use profile and executable failure suite, offered for testing against existing practice. The three cases where no additional layer is needed are an important falsifiability condition, not a concession to remove.

Judgment

Support inclusion in the curated public seed. No unsupported priority claim, standards-conformance assertion, legal-authority transfer or demonstrated adoption was found in the repaired memo. A complete prior-art survey may find closer work; the memo says so. The next substantive test is an independently implemented, permission-respecting adapter with semantic loss and total labor measured—not another vocabulary table presented as interoperability. The final export must be rebuilt to include these new documents and repairs; an earlier ZIP’s successful reproduction does not attest to their bytes.