Rights-Grade Pre-Release Admissibility: Policy Evaluation Before Content Commitment

Mechanism

The gate sits at the boundary between content production and content commitment. Before an artefact may be assigned an anchor and entered into the lineage, it must pass through an evaluation that consults a policy object specific to the destination into which the artefact is being committed. The policy object is itself a committed artefact: it carries a signature from the governance authority, an explicit statement of the lineage requirements, an enumeration of the authority credentials that are accepted, and a definition of the governance classes the destination admits. The gate's role is to determine whether the candidate artefact, together with the credentials presented by the producer, satisfies the policy.

The lineage axis examines the candidate artefact's history. Content that derives from another anchored artefact carries a derivation record showing the parent anchor and the bounded operations applied. The gate verifies that the parent anchor exists, that the chain of derivations leads to an admissible root, and that every bounded operation in the chain falls inside its declared envelope. Content that does not derive from a parent (originally produced content) carries a producer-attestation record instead, and the gate verifies the attestation against the producer's credentials. The lineage axis fails if the chain has gaps, if any bounded operation exceeds its envelope, or if the root is not admissible under the destination's policy.

The authority axis examines the producer's credentials. The producer presents a credential object that asserts the right to publish into the destination's governance class. The gate verifies the credential's signature against the trust roots referenced by the policy and verifies that the credential has not been revoked by any revocation source the policy designates as authoritative. The authority axis fails if the signature does not verify, if the trust path does not connect to a designated root, or if any node in the path appears in a revocation source. Authority verification is structural rather than reputational; it does not consult historical behaviour.

The governance-class axis examines the declared category of the content. The producer declares a class at submission, and the gate verifies that the class is one the destination admits and that the declared class is consistent with structural properties of the content. For example, a class that requires the content to be human-readable text is checked against the artefact's structural-variance descriptors to confirm consistency with the textual band set; a class that requires the content to be derived from a recognised generative model is checked against a model-attestation record. The governance-class axis fails if the declaration is unrecognised, if the destination does not admit the declared class, or if the structural check is inconsistent with the declaration.

A submission must pass all three axes to be admitted. There is no compensating mechanism by which a strong showing on one axis can offset a failure on another; the gate is conjunctive. Admission produces an admission record that is committed into the lineage alongside the new anchor; the admission record references the policy version under which the gate evaluated, the credentials verified, the parent anchors consulted, and the governance class assigned. Non-admission produces a refusal record that is also committed (so that the attempt is auditable) but does not produce an anchor; without an anchor, the artefact is not publishable into the destination.

The gate is not a recommendation engine. It does not produce a probability of admissibility or a ranked decision; it produces a deterministic admit-or-refuse outcome, with the outcome and its supporting evidence committed to lineage. This determinism is essential for downstream consumers, who must be able to rely on the admission as a structural property rather than as a heuristic estimate.

Operating Parameters

The policy object is the principal parameter. It is structured as a versioned, signed document that carries a destination identifier, a list of accepted trust roots, a list of revocation sources, an enumeration of admitted governance classes, and a lineage-requirement specification. The lineage requirement may demand fully anchored derivation, may admit producer-attested origination, or may admit both with class-conditional rules. The policy object is itself referenced by a stable identifier from any admission record produced under it, so that historical admission decisions can be re-evaluated against the policy that was in force at the time.

Trust roots are explicit cryptographic anchors against which producer credentials are verified. The policy enumerates the roots; the gate does not consult any root not listed. Trust roots may be rotated through a governance-commit process that itself satisfies the policy then in force, so that root rotation is an auditable structural event rather than an out-of-band administrative action.

Revocation sources define where the gate consults to determine whether a credential has been revoked. A revocation source may be a credential-status list signed by a designated authority, or it may be a structural revocation log maintained as part of the lineage system. The policy specifies the freshness requirement for each source: a source whose latest signed update is older than the freshness bound is treated as unavailable, and the policy specifies whether unavailability is fail-closed (the gate refuses) or fail-open (the gate proceeds without consulting the source). Fail-closed is the typical setting for high-assurance destinations.

Governance classes are enumerated identifiers that carry, in the policy, the structural checks required for content of that class. The class definition includes the band set against which structural-variance consistency must hold, the bounded-transform envelope that derived content is permitted to occupy, and the producer-credential requirements specific to the class. Classes may be hierarchical: a parent class definition is inherited by children unless explicitly overridden.

Admission and refusal records carry the timestamp of the evaluation, the policy version, the credentials verified, the parent anchors consulted, and (in the case of refusal) the axis or axes on which the evaluation failed. The records are committed into the lineage and become part of the audit surface; they cannot be retroactively altered without breaking the cryptographic chain.

Alternative Embodiments

Several embodiments are contemplated. In a single-destination embodiment, every gate evaluates against a single policy object, and the producer interacts with one set of governance classes. This is appropriate for tightly scoped deployments such as a single newsroom or a single broadcaster. In a multi-destination embodiment, the producer specifies the destination at submission, and the gate selects the appropriate policy object; a single producer may publish into multiple destinations with different admissibility rules, and the lineage records which destination admitted which artefact.

In a federated embodiment, multiple destinations cross-recognise each other's admission records under defined conditions. Federation reduces friction for content that flows between cooperating destinations; it does not relax admission. Each destination still evaluates its own policy, but a candidate that carries an admission record from a federated peer may have part of its lineage axis pre-validated. The federated embodiment is governed by a federation policy that itself satisfies the same structural commitments as a destination policy.

In a delegated embodiment, the gate may delegate part of its evaluation to a specialised verifier. For example, the structural consistency check on the governance-class axis may be delegated to a service that holds expertise in a specific medium. The delegation is itself an admission-record event, so the delegate's outcome is auditable, and the destination retains responsibility for the overall admission decision.

A pre-flight embodiment permits producers to submit candidate artefacts to a non-committal evaluation before formally requesting admission. The pre-flight returns the same outcome the gate would produce but does not commit a record. This is useful for workflow tooling that wants to surface admissibility issues during editing rather than at publication time. The pre-flight is structurally distinct from admission: it produces no anchor, no lineage entry, and no permission to publish.

A class-conditional embodiment varies the gate's strictness with the declared governance class. Low-stakes classes may admit producer-attested origination without requiring full anchored lineage; high-stakes classes may require complete anchored lineage to a recognised root. The class-conditional logic is encoded in the policy and is therefore subject to the same governance-commit machinery as any other policy parameter.

Composition

Pre-release admissibility composes with the anchoring substrate as the gate that controls what enters the substrate. Without admission, no anchor is issued; without an anchor, the content cannot be verified at downstream consumers and cannot accumulate lineage. The two primitives are sequentially dependent: admissibility is the necessary prelude to anchoring, and anchoring is the necessary prelude to verifiability.

Composition with adversarial robustness is tight. Admitted content acquires anchors that include the cross-band descriptors and the bounded-transform manifest used by the robustness primitive. Non-admitted content has no anchor and therefore registers as unanchored at any downstream verification, falling into the third outcome class (unverified) rather than the failure class. The two primitives jointly produce the three-outcome surface: admitted-and-corroborating, admitted-and-failing, never-admitted.

Composition with lineage produces the audit story. Every admission record, every refusal record, and every policy version is a committed lineage entry. A regulator or auditor can reconstruct the full population of artefacts that were admitted into a destination over any time window, the policy in force during that window, the credentials presented, and the outcomes. This reconstruction is performed offline against the lineage; it does not require the destination's operational systems to be available.

Composition with policy distribution is what permits the gate to operate without a centralised admission service. Because the policy object is itself committed and cryptographically signed, every gate instance operates from the same policy regardless of where the gate executes. A federated deployment may have many gate instances running across many substrates, and they will produce identical decisions on identical inputs because they consult the same committed policy.

Composition with credential issuance closes the loop. Credentials are themselves anchored artefacts produced through their own admissibility process; the credential-issuance policy gates which producers may obtain which classes of credentials, and credentials are committed into a credentialing lineage that the gate consults at admission time. The recursion bottoms out at the trust roots, which are committed once at deployment and rotated through governance commits.

Prior-Art Differentiation

Conventional content moderation systems evaluate content after publication, removing or hiding artefacts that violate policy. Post-hoc moderation is incompatible with rights-grade governance because it depends on detection at scale rather than on prevention by construction. The present mechanism is structurally different: an artefact that fails the gate cannot publish in the first place, so the moderation problem is replaced by an admission problem that operates on a smaller surface and produces auditable records.

Workflow-based publication systems gate publication on internal review steps, but the gate is typically procedural rather than structural: an editor clicks "approve," and the artefact becomes published. Procedural gates are subject to bypass through error or compromise, and they do not produce verifiable evidence that any particular policy was applied. The present mechanism replaces procedural approval with structural admission and produces a committed admission record that downstream consumers can verify.

Digital-rights-management systems control consumption rather than admission; they assume that content has already been published and seek to constrain who may decrypt or display it. The present mechanism operates upstream, at the moment content enters the publishable substrate, and is concerned with whether the content acquires the structural identity that makes consumption possible. The two approaches address different failure modes and may be combined.

Content-licensing registries record permissions but typically do not gate the production of the content itself; they are consulted at distribution time and rely on participants to comply. The present mechanism is enforcement by construction: the gate is the only path through which content acquires an anchor, and content without an anchor cannot publish, so non-compliance does not produce a slow-moving infringement problem but is structurally prevented.

The closest prior art in policy-evaluation gates is found in software-supply-chain admission controllers, which evaluate signed policy against candidate artefacts before allowing them to be deployed. The novel contribution here is the application to content rather than to executable artefacts, the three-axis evaluation that combines lineage, authority, and governance class, and the integration with the anchoring substrate so that admitted content acquires a structural identity rather than a deployment-time approval flag.

Disclosure Scope

The disclosure encompasses the three-axis admission gate, the conjunctive evaluation rule, the committed policy object, the admission and refusal records as committed lineage events, and the gate's role as the necessary prelude to anchoring. It encompasses the single-destination, multi-destination, federated, delegated, pre-flight, and class-conditional embodiments described above, and the parameter set that defines trust roots, revocation sources, freshness bounds, and governance-class definitions.

The disclosure does not constrain the specific cryptographic primitives used for signatures, the specific structural-variance descriptors used for class consistency checks, or the specific governance-class identifiers used in any deployment. These are deployment choices that depend on the application and the threat model. The disclosure does constrain the structural properties: the gate must be conjunctive across the three axes, the policy must be committed and versioned, the admission and refusal outcomes must be recorded as lineage events, and the gate must be the only path by which content acquires an anchor.

The disclosure is filed under Provisional 63/808,372 and constitutes a structural primitive of the content anchoring system. A system that publishes without gating, or that gates without committing the policy version, or that admits content on a non-conjunctive (compensatory) rule, is outside the disclosure. The structural properties are jointly necessary; no proper subset is sufficient.