Skip to content

OPENETR/CD 1:2026

Evidence Graphs - Open protocol requirements for end-verifiable evidence and consequential state

Committee draft

Status: Draft for discussion and technical review
Edition: 1
Date: 2026-09-10 Language: English working draft

Important notice: This document is an OpenETR project draft prepared in the style of a consensus standard. It is not a National Standard of Canada, has not been developed or approved by an organization accredited by the Standards Council of Canada, and has not been approved or endorsed by the Standards Council of Canada. The identifier above is a project committee draft identifier only and is not a Canadian standards designation.

Preface

This committee draft specifies a general model for identifying a Digital Artifact, preserving end-verifiable evidence concerning it, and deriving Consequential State from that evidence under an identified policy.

The draft is intended to support interoperability among independent systems, organizations, sectors, and legal jurisdictions. It does not prescribe a single application, registry, database, relay operator, document repository, identity provider, trust framework, or legal rule book.

The initial interoperable wire binding uses signed Nostr events. The core model is transport-neutral so that another event protocol or storage system can implement an equivalent evidence model if it preserves the required properties.

This draft should be reviewed through an open, balanced, consensus-oriented standards-development process before it is considered for adoption by any standards body. Such a process should include representation from users, implementers, public-interest participants, legal and policy experts, regulated sectors, conformity-assessment interests, and relevant authorities.

Introduction

Digital content can be copied perfectly. A copy of a file can reproduce its bytes, but the file alone cannot establish who made a consequential statement about it, how that statement relates to prior statements, or what state follows under defined rules.

OpenETR separates five concerns:

  1. the Digital Artifact, identified by its content;
  2. the Evidence Graph, containing signed evidence;
  3. Consequential State, derived from validated Evidence Graph records;
  4. Recognition, supplied by an applicable external rule book; and
  5. Effect, given to recognized state by a person, system, institution, contract, or law.

The model may be summarized as follows:

Digital Artifact -> Evidence Graph -> Consequential State

Digital Artifact + established Consequential State -> Digital Original

Consequential State -> Recognition -> Effect

OpenETR does not make an application database the permanent source of truth. Applications may store projections and indexes, but a conforming verifier can reconstruct the evidence and derive state independently.

1 Scope

1.1 General

This document specifies requirements and recommendations for:

a) identifying a Digital Artifact by cryptographic digest;

b) constructing an Evidence Graph from signed, end-verifiable records;

c) anchoring and extending a candidate Evidence Graph, including any Control Graph subset;

d) verifying record integrity, signer attribution, object consistency, and graph links;

e) deriving Consequential State under an identified verifier policy;

f) representing generic control actions;

g) separating cryptographic verification from recognition and effect;

h) representing actor-neutral signing identities;

i) integrating OpenETR into independent host systems and domain adapters;

j) storing and retrieving evidence without dependence on one application at the time of performance; and

k) implementing the OpenETR Nostr wire profile.

1.2 Application

This document is applicable to durable electronic records and other Digital Artifacts for which independently verifiable consequential history is useful. Examples include warehouse receipts, bills of lading, product passports, certificates, credentials, permits, health records, provenance records, and other electronic records.

1.3 Exclusions

This document does not determine:

a) legal title, ownership, negotiability, possession, priority, or protected holder status;

b) whether an electronic record satisfies legislation in a jurisdiction;

c) KYC, AML, sanctions, licensing, registry, or accreditation sufficiency;

d) the authority of a signer to act for a person or organization;

e) the substantive content required for a domain document;

f) document custody, confidentiality, disclosure, or retention obligations;

g) whether a relying party is required to recognize a record; or

h) the legal, contractual, regulatory, or institutional effect of a record or state.

2 Normative references

The following documents are referenced by the Nostr conformance profile. For undated references, the implementing conformance statement shall identify the revision or commit used.

  • NIST FIPS PUB 180-4, Secure Hash Standard (SHS).
  • Nostr NIP-01, Basic protocol flow description.
  • Nostr NIP-19, bech32-encoded entities.
  • OpenETR, OpenETR Nostr Wire Format Specification.
  • OpenETR, Evidence Event Minimum Shapes.

3 Terms and definitions

For the purposes of this document, the following terms and definitions apply.

3.1 actor

person, organization, service, autonomous agent, signing service, or combined workflow associated by an integrating system with a signing key

Note 1 to entry: The protocol does not infer actor type from a public key.

3.2 Anchor Record

first signed record in a candidate Evidence Graph concerning a Digital Artifact

Note 1 to entry: An Anchor Record supplies a candidate starting point. It does not by itself establish uniqueness, recognition, or effect.

3.3 candidate chain

ordered path through cryptographically linked records that has not necessarily been accepted under the selected verifier policy

3.4 Commitment Profile

independent operational signing identity organized by a Control Desk Key and used to sign OpenETR records

3.5 Consequential Action

action evidenced by a signed record that can affect derived state under an applicable policy

3.6 Consequential State

state that follows when validated Evidence Graph records are evaluated according to defined protocol and verifier rules

Note 1 to entry: Consequential State is a derived result, not another event or authoritative database row.

3.7 Contact

locally named external party whose public key can be resolved for interaction

Note 1 to entry: A Contact name does not establish authority, familiarity, or recognition.

3.8 control graph

subset of an Evidence Graph formed by the records to which defined rules assign controller or control-transition consequences concerning the same Digital Artifact

3.9 evidence record

signed, end-verifiable record that provides evidence of an action, assertion, or relationship concerning a Digital Artifact and participates in an Evidence Graph

Note 1 to entry: An evidence record does not necessarily change control or Consequential State. Defined rules determine what consequence, if any, follows.

3.10 Control Desk Key

administrative root signing identity used by an OpenETR environment to manage configuration and Commitment Profiles

Note 1 to entry: The Control Desk Key and Commitment Profile keys are ordinary, cryptographically independent key pairs. Their administrative relationship is created by the OpenETR component, not by the underlying signature algorithm.

3.11 Digital Artifact

persistent digital content with a unique content identity, normally established by a cryptographic digest

3.12 Evidence Graph

signed evidence structure from which defined rules derive Consequential State concerning a Digital Artifact after the evidence is validated

Note 1 to entry: An Evidence Graph may consist of a single Anchor Record or linked end-verifiable records, including later assertions and supporting evidence. The Control Graph is the subset to which a ruleset assigns control consequences.

Note 2 to entry: Evidence Graph replaces the former term Digital Controllable Record (DCR). This terminology change does not alter event kinds, signatures, digest identification, or existing wire-format fields.

3.13 Digital Original

Digital Artifact with Consequential State

Note 1 to entry: Consequential State is established through an Evidence Graph. Recognition and effect remain external determinations.

3.14 domain adapter

component that maps domain vocabulary, roles, actions, and policies to and from the generic OpenETR model

3.15 effect

legal, contractual, regulatory, institutional, operational, or other consequence given to recognized state

3.16 end-verifiable

capable of being verified by a relying party from the evidence itself without requiring the application that created or displayed the evidence to be running

3.17 event

cryptographically signed data structure containing an identifier, signer, timestamp, type, tags, content, and signature

3.18 Evidence Event

cryptographically signed event that serves as an evidence record in an Evidence Graph

Note 1 to entry: In the OpenETR Nostr binding, a kind 1415 Anchor Event and a kind 1416 event are Evidence Events. The Anchor Event begins a candidate Evidence Graph; later Evidence Events extend or relate to it.

Note 2 to entry: The existence of an Evidence Event does not by itself establish a state change, recognition, or effect.

3.19 evidence graph

See 3.12. This clause is retained as a cross-reference to preserve existing clause numbering; it does not define a second construct.

3.20 guard policy

policy applied before publication by an implementation to decide whether it will sign or publish a proposed event

Note 1 to entry: A guard policy cannot prevent another implementation from publishing a competing signed event.

3.21 Key-Based Identifier

KBI

identifier whose canonical value is public-key verification material or a deterministic encoding of that material

Note 1 to entry: A KBI identifies the signing key used to verify attributable signed evidence. It does not, by itself, establish the identity, actor type, authority, role, or recognition of the actor associated with that key.

3.22 recognized state

Consequential State accepted under an identified external recognition policy for a stated purpose

3.23 recognition

determination by a relying party, institution, registry, trust framework, contract, or law that evidence or derived state is acceptable for a stated purpose

3.24 Reference

external authority, registry, assessor, attestor, trust service, or other source that can supply recognition or assurance context

3.25 relying party

person, organization, system, or authority that evaluates OpenETR evidence or acts upon resulting state

3.26 verifier policy

identified and versioned rule book used to validate Evidence Graph records, enumerate candidate chains, issue findings, and derive Consequential State

3.27 warning

structured finding that identifies a policy or evidence concern without making otherwise authentic signed evidence disappear

4 Abbreviated terms

AML anti-money laundering
API application programming interface
CLI command-line interface
ETR electronic transferable record
JSON JavaScript Object Notation
KBI Key-Based Identifier
KYC know your customer
MLWR Model Law on Warehouse Receipts
NIP Nostr Implementation Possibility
REST representational state transfer
SHA-256 Secure Hash Algorithm with a 256-bit output

5 Conventions and conformance

5.1 Verbal forms

In this document:

  • shall indicates a requirement;
  • should indicates a recommendation;
  • may indicates a permission; and
  • can indicates a possibility or capability.

5.2 Conformance targets

An implementation may claim conformance as one or more of the following:

a) Core Publisher, which creates signed OpenETR records;

b) Core Verifier, which verifies evidence and derives state;

c) Domain Adapter, which maps a domain to the OpenETR core;

d) Nostr Publisher, which implements the normative Nostr profile in Annex A;

e) Nostr Verifier, which retrieves and verifies the normative Nostr profile in Annex A; or

f) Integration Surface, which exposes the component through a machine or human interface.

5.3 Conformance statement

A conformance claim shall identify:

a) the conformance target or targets;

b) the version of this document;

c) the implemented wire binding and its version;

d) the digest and signature algorithms;

e) the implemented action set;

f) the verifier policy identifier and version, where state is derived;

g) any domain profile implemented;

h) supported identifier encodings; and

i) known deviations, extensions, or unsupported optional features.

6 OpenETR reference model

6.1 Separation of concerns

A conforming implementation shall distinguish:

a) Digital Artifact content;

b) signed Evidence Graph records;

c) derived Consequential State;

d) recognition decisions; and

e) external effect.

An implementation shall not represent a mutable application state projection as if it were the signed Evidence Graph records that produced it.

6.2 Layer model

The OpenETR model consists of four logical layers:

Domain adapter
  domain vocabulary, workflows, validation, and policy presentation

OpenETR protocol layer
  Evidence Graph records, graph traversal, state transition rules, and state derivation

Wire binding
  signed records, identifiers, tags, transport, storage, and retrieval

Recognition layer
  law, contracts, registries, trust frameworks, and relying-party decisions

A conforming implementation may combine these layers in one deployment, but shall preserve their logical boundaries.

6.3 Technology neutrality

The core model shall not require Nostr where an alternative binding preserves:

a) cryptographic attribution to a KBI;

b) stable artifact commitment;

c) immutable record identity;

d) record signature;

e) signed structured data;

f) explicit graph links;

g) replayable evidence history; and

h) independent verifier-policy inputs.

Implementations that exchange records using the OpenETR Nostr profile shall conform to Annex A.

7 Digital Artifact identification

7.1 Digest identity

A Core Publisher shall identify a Digital Artifact by a cryptographic digest.

The digest algorithm shall provide pre-image resistance, second-pre-image resistance, and collision resistance appropriate to the application.

The OpenETR Nostr profile shall use SHA-256.

7.2 Content boundary

The domain adapter or publisher shall define the bytes or canonical data over which the digest is calculated.

Where canonicalization is used, the implementation shall document:

a) the canonicalization algorithm and version;

b) package finalization rules;

c) treatment of metadata and signatures embedded in the artifact;

d) whether transformations produce a new Digital Artifact; and

e) how alternate renderings relate to the canonical content.

7.3 Copying

Byte-identical copies having the same digest shall be treated as instances of the same Digital Artifact.

Copying artifact bytes or Evidence Graph records shall not be interpreted as creating a separate Digital Artifact, a new Evidence Graph history, another controller, or independent Consequential State.

A copy of complete Evidence Graph records may be used by another conforming verifier to reconstruct the same candidate graph and derive the same Consequential State under the same identified rules.

7.4 Disclosure

An implementation may disclose the artifact, retain it privately, or make it available through a separate content service. The Evidence Graph shall not require the artifact bytes to be stored inside each evidence record.

8 Evidence Graph requirements

8.1 General

An Evidence Graph shall consist of one signed Anchor Record or a graph beginning with an Anchor Record and extended by signed records concerning the same Digital Artifact.

Each record shall be independently verifiable for integrity and signer attribution.

8.2 Anchor Record

An Anchor Record shall:

a) identify the Digital Artifact;

b) identify its signer;

c) have a cryptographically derived record identifier;

d) carry a valid signature; and

e) declare an anchoring or issuance action.

An Anchor Record shall be treated as a candidate assertion. A verifier shall not infer universal authority or uniqueness merely because an Anchor Record is cryptographically valid.

More than one candidate Anchor Record may exist for the same Digital Artifact.

8.3 Later evidence records

A later evidence record shall:

a) identify the same Digital Artifact as the candidate Anchor Record;

b) identify the immediately prior record or specific record on which it relies;

c) identify its action;

d) identify its signer; and

e) carry a valid signature.

An evidence record shall not be silently relinked to a different prior record.

8.4 Generic actions

The generic OpenETR action vocabulary includes:

  • issue, to create an Anchor Record;
  • initiate, to propose a transfer;
  • accept, to accept a proposed transfer;
  • attest, to make a signed assertion;
  • notice, to state the Anchor Publisher's position concerning an artifact;
  • encumber, to declare a claimed encumbrance;
  • discharge, to release or satisfy a referenced encumbrance;
  • redeem, to present the artifact to an obligor for performance; and
  • terminate, to end the active lifecycle if recognized under policy.

A domain adapter may assign domain terminology to these actions. It shall preserve the generic action semantics when claiming OpenETR interoperability.

8.5 Structured and unstructured data

Machine-readable event data shall be represented as signed structured fields or tags.

Human-readable comments, summaries, and other unstructured context should be represented separately from required structured data.

A verifier shall not be required to parse human-readable content to recover a required graph link, artifact identifier, participant, action, or action-specific reference.

8.6 Linked evidence

An implementation may associate supporting artifacts or attestations with an Evidence Graph. Linked evidence shall identify:

a) the primary Digital Artifact or a specific Evidence Graph record;

b) the linked artifact or evidence item;

c) the signer making the association; and

d) the claimed relationship type.

Linked evidence shall not be interpreted as changing controller state unless the selected policy explicitly assigns that consequence.

A linked-evidence relationship shall distinguish an evidentiary relationship from a relationship to which defined rules assign a state consequence. Evidentiary relationships may include derived_from, ingredient_of, and rendition_of. These relationships shall not, by themselves, be interpreted as meaning that one artifact supersedes or replaces another.

Relationships such as supersedes and replaces shall affect Consequential State only where the selected verifier policy or Domain Adapter defines their preconditions and effects.

Where linked evidence depends on an external verification mechanism, verifier output should distinguish, where applicable, evidence availability, integrity, content binding, signature or credential status, policy sufficiency, and recognition. Provenance evidence shall not be represented as proving control, supersession, recognition, or effect unless separate defined rules support that conclusion.

8.7 Attestations

An attestation should identify the specific event, graph commitment, evidence bundle, observation, or procedure to which it applies.

A valid attestation signature shall establish attribution of the attestation. It shall not, by itself, establish that the attestor is:

a) the controller of the Digital Artifact;

b) the author of the underlying event;

c) the source of Consequential State;

d) a canonical registry; or

e) a recognition authority.

Defined rules shall determine what an attestation contributes to Consequential State. An external recognition context shall determine what effect, if any, to give the attestation or attestor.

9 Graph reconstruction and verification

9.1 Retrieval

A Core Verifier shall be capable of retrieving or accepting candidate records for a supplied artifact identifier from one or more identified evidence sources.

The verifier should avoid assuming that the first record returned by one repository is authoritative.

The verifier shall not represent records returned by declared evidence sources as proving that no additional or conflicting record exists elsewhere.

9.2 Structural verification

For each candidate record, a Core Verifier shall verify:

a) record identifier integrity;

b) signature validity;

c) signer identifier syntax;

d) artifact identifier syntax;

e) required record fields;

f) consistency of the artifact identifier across a candidate chain;

g) existence and identity of referenced prior records; and

h) action-specific required data.

9.3 Graph enumeration

A Core Verifier shall enumerate candidate Anchor Records and candidate graph paths before applying recognition conclusions.

The verifier shall preserve competing, branching, or policy-questionable signed evidence for inspection.

A record whose required prior-record link is absent, malformed, cyclic, or cryptographically inconsistent shall not be used as a valid extension of that candidate chain.

The verifier should report the affected record and reason.

9.5 Reproducibility

Given the same evidence set, policy identifier, policy version, and evaluation parameters, conforming verifiers shall produce equivalent structural findings and Consequential State.

9.6 Verification result dimensions

A Core Verifier shall not represent all verification and recognition concerns as one undifferentiated validity result.

Where applicable, verifier output shall distinguish:

a) artifact integrity;

b) event authenticity;

c) structural validity;

d) graph continuity;

e) historical continuity, where evaluated;

f) source consistency or divergence, where observations can be compared;

g) transition validity;

h) derived Consequential State;

i) retrieval coverage;

j) evidence sufficiency under the identified policy;

k) optional temporal assurance;

l) observation, witness, or audit attestation status;

m) linked-evidence and external-verification status, including provenance where evaluated;

n) monitoring status, where applicable;

o) actor or authority recognition;

p) system reliability assessment; and

q) external recognition and effect.

An evaluated dimension should use a defined outcome vocabulary that can distinguish valid, invalid, unverifiable, conflicting, absent, not_evaluated, and not_applicable where those outcomes are meaningful.

An outcome for one dimension shall not be silently treated as an outcome for another dimension.

9.7 Divergent observations and equivocation

Valid signatures shall not be represented as proving that every observer has received a single globally consistent Evidence Graph.

Where observations from multiple evidence sources can be compared, a verifier should report stale, incomplete, incompatible, or divergent evidence sets. If the available evidence and rules do not support a unique state, the verifier shall preserve the competing evidence and report the state as ambiguous, conflicting, insufficiently evidenced, or otherwise unresolved.

The result shall identify the evidence sources, retrieval scope, rules, and evaluation parameters on which the conclusion depends.

10 Consequential State

10.1 Derivation

Consequential State shall be derived from validated Evidence Graph records under an identified verifier policy.

An implementation shall be able to identify the evidence records and policy version that produced a displayed state.

10.2 State projection

An application may cache or index Consequential State. A cached projection shall not replace the underlying evidence and shall be invalidated or recalculated when the relevant evidence set or policy changes.

10.3 Generic state dimensions

A verifier may derive:

a) lifecycle state, including active, transfer pending, redemption pending, or terminated;

b) candidate current controller;

c) candidate transfer participants;

d) declared encumbrances, discharges, and outstanding encumbrances;

e) attestations and linked evidence; and

f) structural and policy findings.

10.4 Digital Original

A Digital Original shall have Consequential State established through an Evidence Graph.

An implementation shall not claim that byte possession alone, file location, or an application database row establishes a Digital Original.

Whether the Digital Original and its state are recognized as authoritative for a particular purpose is determined under Clause 13.

10.5 Historical state derivation

Where sufficient evidence is available, a Core Verifier should be capable of deriving Consequential State for a candidate Evidence Graph path ending at a specified event.

A historical-state result shall identify:

a) the terminal event or graph cut;

b) the ancestor evidence used;

c) the evidence sources and retrieval scope;

d) the verifier policy identifier and version;

e) competing evidence present in the supplied evidence set;

f) unresolved links or completeness limitations; and

g) the resulting Consequential State and findings.

The result shall be represented as evidence-set-relative, rule-set-relative, and path-relative. It shall not be represented as globally authoritative state at the event's declared timestamp unless separate accepted evidence supports that conclusion.

11 Verifier policy

11.1 Policy identification

A verifier policy used to derive state shall have a stable identifier and version.

The verifier output shall identify the policy applied.

11.2 Baseline evaluation

A baseline verifier policy shall:

a) retrieve or accept the artifact evidence set;

b) verify cryptographic and structural correctness;

c) enumerate candidate chains;

d) evaluate generic transition rules;

e) identify ambiguity and conflicts;

f) produce structured findings; and

g) derive candidate Consequential State.

A verifier policy claiming OpenETR baseline-policy conformance shall publish normative transition rules for every action from which it derives Consequential State. For each supported action, those rules shall define:

a) permitted parent record kinds and actions;

b) required signer and participant relationships;

c) required structured evidence;

d) state preconditions;

e) state variables created, changed, or cleared;

f) handling of duplicate, competing, missing, or out-of-order evidence;

g) the conditions that produce invalid, conflicting, unverifiable, warning, or non-recognition outcomes; and

h) treatment of unknown actions and extensions.

The policy shall also define whether and how an Anchor Record establishes initial candidate controller, lifecycle, or other state. Where the applicable rules do not define a consequence, a verifier shall not infer one from an action label, signer identity, document text, or application state.

11.3 Evidence preservation

A policy violation shall not cause an otherwise authentic signed record to be erased from the evidence set.

A verifier shall distinguish:

a) evidence that is cryptographically or structurally unusable;

b) evidence that is authentic but contrary to a policy rule; and

c) evidence accepted as effective under the selected policy.

11.4 Warnings

Policy concerns should be returned as structured warnings or non-recognition annotations unless the concern prevents structural verification.

A warning should include:

a) a stable machine-readable code;

b) severity;

c) the affected record or identity;

d) a human-readable explanation; and

e) the policy consequence.

11.5 Domain and organizational policies

A domain or organization may define a policy overlay that adds safeguards, actor requirements, schemas, attestations, exemptions, or recognition rules.

An overlay shall not conceal the baseline signed evidence or structural result.

Different relying parties may reach different recognition conclusions from the same authentic evidence set. Each conclusion shall identify the policy under which it was reached.

12 Identity and key organization

12.1 Actor neutrality

The OpenETR protocol shall treat a signing key independently of whether it is operated by a human, organization, service, autonomous agent, hardware device, or combined workflow.

A valid signature shall establish that the signed record was produced using the private-key capability corresponding to the identified KBI. It shall not be represented as proof of key custody, actor intent, actor type, legal identity, mandate, licensing, KYC status, or authority.

12.2 Principal, operator, and signer

An integrating system should distinguish:

a) the principal, on whose authority an action is performed;

b) the operator, which initiates or executes the action; and

c) the signer, which produces the cryptographic signature.

The system should retain or link evidence of these relationships where its policy requires them.

12.3 Control Desk Key and Commitment Profiles

A Control Desk Key may organize one or more Commitment Profiles.

The relationship shall be treated as an application-level administrative relationship, not as cryptographic derivation or inherent delegation between the keys.

The same key may perform both roles, but production implementations should separate administrative and operational signing roles where practical.

12.4 Acting Profile

Where more than one Commitment Profile is available, the user or host system shall be able to determine which Commitment Profile is the Acting Profile for a proposed operation.

A signed operational record shall identify the public key that actually signed it, regardless of the profile label displayed by the application.

12.5 Contacts and References

A Contact may resolve a local name to an external participant public key.

A Reference may identify an external recognition or assurance source.

Neither a Contact nor a Reference shall, by itself, be treated as proof of legal identity, authority, or trustworthiness.

12.6 Identifier resolution

Human-facing interfaces may accept profile names, Contact names, public-key encodings, hexadecimal keys, or externally resolvable identifiers.

Before signing or comparison, an implementation shall normalize an identity to the canonical public-key representation required by its wire binding.

Identifier resolution shall not be represented as recognition.

12.7 External key-history evidence

An implementation may consume or link evidence concerning historical or current bindings between an actor identifier and a signing key. Such evidence may come from a Key Transparency system, PKI, DID method, credential system, trust registry, or another external mechanism.

Where later historical verification depends on external evidence, the implementation should preserve or reference the evidence used at evaluation time. A current directory response shall not be represented as proof of a historical key binding.

External key-history evidence shall be evaluated under its own verification and recognition rules. OpenETR shall not require one key-history mechanism.

12.8 Key protection

Private keys shall not be published in OpenETR events, logs, API responses, or machine-readable command output.

An implementation shall protect private keys according to the risk and assurance requirements of the host system. It should support separation of administrative and operational keys, backup, rotation, recovery, suspension, and incident response.

13 Recognition and effect

13.1 Boundary

OpenETR verification shall answer which key signed which record concerning which Digital Artifact and how the record participates in a candidate Evidence Graph.

OpenETR core shall not claim that cryptographic validity alone establishes legal, institutional, contractual, or regulatory effect.

13.2 Recognition inputs

A verifier policy may use recognition inputs including:

a) trust registries;

b) KYC or identity providers;

c) licensing or regulatory registries;

d) attestations and credentials;

e) Web of Trust signals;

f) organizational familiarity lists;

g) enterprise account and role systems;

h) contractual network rules; and

i) applicable legislation or judicial determinations.

13.3 KYC and actor assurance

KYC, AML, sanctions, agent governance, and similar actor-assurance controls shall be treated as integration or recognition concerns unless a domain profile expressly requires them.

Their absence shall not be represented as invalidating an otherwise valid event signature.

13.4 Transparent conclusions

A recognized-state output should distinguish:

a) cryptographic findings;

b) structural graph findings;

c) protocol-policy findings;

d) recognition inputs consulted; and

e) the resulting effect or recommendation.

14 Domain adapters

14.1 General

A Domain Adapter shall translate domain actions and terms to the generic OpenETR model and translate Evidence Graph results back into domain language.

14.2 Required profile information

A conforming Domain Adapter shall document:

a) scope and excluded record types;

b) intended users and relying parties;

c) Digital Artifact boundaries and canonicalization;

d) digest algorithm;

e) domain roles and their OpenETR mappings;

f) domain actions and their generic action mappings;

g) required structured event data;

h) verifier policy and recognition inputs;

i) privacy, disclosure, and retention assumptions;

j) error and warning presentation; and

k) any jurisdiction-specific requirements.

A Domain Adapter may use legal or business terminology appropriate to its users. It shall not imply that the generic OpenETR term itself satisfies a legal classification.

For example, an OpenETR Evidence Graph is not automatically a controllable electronic record under UCC Article 12 or an electronic transferable record under an MLETR-based enactment.

14.4 Document independence

A Domain Adapter shall permit the protocol layer to remain logically separate from document movement and storage.

The artifact may be exchanged by an existing document-management, registry, messaging, repository, or API system.

15 Storage, transport, and availability

15.1 Independence from a running application

A conforming system shall enable signed Evidence Graph records to be exported or retrieved in a form that can be verified without the originating application being available at the time of performance.

15.2 Repository neutrality

Evidence Graph records may be stored on public relays, private relays, local relays, databases, archives, files, or other repositories.

Repository acceptance shall not be treated as recognition of an event's semantic or legal effect.

15.3 Redundancy

Implementations should use redundant retrieval or archival arrangements appropriate to the required availability and retention period.

A relying party should be able to preserve the exact signed evidence needed to reproduce a consequential-state determination.

15.4 Local operation

An implementation shall not require use of a third party's running OpenETR application where the relying party possesses the required signed events and verification software.

15.5 Relay-backed configuration

An implementation may store non-secret configuration as signed relay-backed records. Bootstrap relay locations and root private-key material may remain local or be protected by a host account or custody system.

15.6 Retrieval coverage

A verifier shall distinguish repository-specific retrieval completion from global evidence completeness.

Where a repository or transport reports that all matching stored records have been returned, the verifier may report completion for that source and query. It shall not infer that no additional or conflicting record exists in another repository or outside the observed evidence boundary.

A verifier should report whether additional records are known to be available, authentication is required, pagination is incomplete, or retrieval completeness is unknown.

Graph continuity and policy sufficiency shall be evaluated separately from repository retrieval coverage.

16 Integration surfaces

16.1 General

The OpenETR component may be exposed through:

a) an installable software library;

b) a command-line interface;

c) machine-readable JSON input and output;

d) REST or other service APIs;

e) a web application; or

f) direct implementation of a wire binding.

All surfaces within one product should use a common service and policy layer so that equivalent operations produce equivalent signed evidence and findings.

16.2 Machine-readable output

A machine-readable interface shall:

a) produce a documented response envelope;

b) separate success, failure, and warning conditions;

c) expose stable reason and warning codes;

d) avoid requiring parsing of human console prose;

e) preserve access to raw signed evidence where appropriate; and

f) exclude private keys and secret recovery material.

A machine-readable verifier result should identify each evaluated dimension, its status, its evidence basis, and the policy or verification procedure used. Command execution success shall not be represented as equivalent to protocol validity, recognition, system reliability, or legal effect.

16.3 Human interfaces

A human interface should use domain terminology and make the Acting Profile, proposed action, target artifact, warnings, and resulting state visible before or after a consequential operation as appropriate.

16.4 Hypermedia

A web implementation may use a hypermedia-driven architecture. Client-side code is not required for protocol conformance.

16.5 Automated interpretation and agentic systems

An implementation using automated extraction, interpretation, or inference shall distinguish machine-generated conclusions from attributable signed or otherwise independently verified evidence.

An inferred conclusion may be retained or linked as evidence where its source, method, output, and attribution are identified. Its evidentiary contribution shall be determined under the selected verifier policy.

Inference alone shall not be represented as proving issuance, actor identity, authority, authorization, control, Consequential State, recognition, or effect. Records produced through an autonomous agent or combined human-agent workflow shall satisfy the same cryptographic, structural, policy, and recognition requirements as records produced through another workflow.

17 Security, privacy, and operational considerations

17.1 Threat model

An implementation shall document threats relevant to its deployment, including:

a) private-key compromise;

b) unauthorized use of a valid signer;

c) conflicting or deceptive records;

d) divergent, stale, or selectively withheld evidence-source results;

e) digest substitution or canonicalization errors;

f) stale state projections;

g) metadata correlation and privacy leakage;

h) denial of service; and

i) dependency or verifier compromise.

An implementation should also consider unauthorized or unexpected use of a cryptographically valid signing key. Signature validity does not establish actor intent.

17.2 Open publication

An implementation shall assume that events published to public relays can be widely replicated and difficult to remove.

Sensitive personal, commercial, health, or confidential data should not be placed directly in public event tags or content unless disclosure is intended and permitted.

17.3 Digest limitations

A digest proves equality to supplied bytes; it does not prove the truth, quality, legality, completeness, or confidentiality of the artifact.

Low-entropy or guessable hidden data may be vulnerable to dictionary attacks even where only a digest is published.

17.4 Algorithm agility

Implementations should support an orderly transition to new digest or signature algorithms. A conformance statement shall identify the algorithms used.

17.5 Time

A signed event timestamp shall not be treated as independently trusted time unless supported by an accepted timestamping or other time-assurance source.

The term Anchor Event is reserved in this document for the initial OpenETR record that begins a candidate Evidence Graph. External timestamp, transparency-log, or ledger evidence should be described as a Temporal Proof or another term that does not imply that it creates OpenETR Consequential State.

Failure to obtain or verify optional Temporal Proof evidence shall not, by itself, invalidate an otherwise valid Evidence Graph. The verifier shall report the temporal-proof outcome separately.

17.6 Auditing and monitoring

An implementation may use independent witnesses, auditors, source comparison, or monitoring services to strengthen evidence about observation and historical continuity.

Auditing and monitoring shall not be represented as establishing controller status, actor intent, recognition, or legal effect unless separate defined rules and recognition inputs support those conclusions.

The OpenETR core shall not require a transparency log, globally consistent event order, auditor network, or monitoring service.

17.7 Post-quantum considerations

The current Nostr profile is not post-quantum secure. Long-lived deployments should preserve algorithm identifiers and migration evidence sufficient to support future re-attestation or transition policies.

18 Conformity assessment

18.1 Core Publisher tests

A Core Publisher conformance assessment shall confirm that the publisher:

a) calculates the declared artifact digest correctly;

b) constructs required structured data;

c) creates stable record identifiers;

d) signs records with the declared signer;

e) links later records to the intended prior record; and

f) does not disclose private keys.

18.2 Core Verifier tests

A Core Verifier conformance assessment shall include:

a) valid and invalid signatures;

b) valid and invalid record identifiers;

c) missing required fields;

d) inconsistent artifact identifiers;

e) missing, cyclic, and conflicting graph links;

f) multiple candidate Anchor Records;

g) competing branches;

h) policy-valid and policy-questionable transitions;

i) reproducible state derivation;

j) historical state derivation at an identified event;

k) divergent or incomplete evidence-source results; and

l) preservation and reporting of authentic but unrecognized evidence;

m) distinct reporting of conflicting evidence; and

n) equivalent results from the same evidence set, policy version, and evaluation parameters.

18.3 Domain Adapter tests

A Domain Adapter conformance assessment shall confirm the documented mappings for artifact scope, roles, actions, structured data, warnings, policy, and recognition boundaries.

Where a Domain Adapter uses automated interpretation, the assessment shall confirm that inferred values are distinguishable from signed or independently verified evidence and are not represented as establishing Consequential State without defined rules.

18.4 Interoperability tests

Two implementations claiming the same wire profile should demonstrate that:

a) one can publish records retrieved by the other;

b) both verify the same record identifiers and signatures;

c) both reconstruct equivalent candidate graph paths; and

d) both derive equivalent state when supplied with the same evidence and policy version.

Annex A (normative) - OpenETR Nostr conformance profile

A.1 General

An implementation claiming OpenETR Nostr conformance shall implement Nostr event serialization, identifier derivation, Schnorr signatures, and relay messages in accordance with NIP-01.

Human-facing bech32 encodings should conform to NIP-19. Canonical event fields and filters shall use the hexadecimal forms required by NIP-01.

In this binding, the 32-byte Nostr public key is the KBI. Its 64-character lowercase hexadecimal representation is the canonical wire encoding. npub is the NIP-19 human-readable encoding of the same KBI and shall not replace the hexadecimal value in NIP-01 event fields or relay filters.

A.2 Event kinds

The profile defines:

Kind Name Requirement
1415 Anchor Event Regular event beginning a candidate Evidence Graph
1416 Evidence Event Regular event extending or relating to a candidate Evidence Graph

OpenETR Evidence Graphs shall use regular events. Addressable or replaceable event kinds shall not be used for Anchor Events or Evidence Events.

A.3 Common tags

The following tag rules apply:

a) o shall contain the 64-character lowercase hexadecimal SHA-256 digest of the Digital Artifact;

b) e shall contain the hexadecimal event identifier of the prior or specific record being referenced;

c) p, where required, shall contain a participant public key in hexadecimal form;

d) action shall identify the semantic action;

e) enc, where required, shall identify the encumbrance event being discharged;

f) type may contain an action-specific subtype; and

g) ref may contain an external or business reference.

Single-letter tags may be used as relay query anchors. Named tags shall be treated as signed structured data even where relays do not index them.

No separate origin or anchor root-pointer tag is required. A verifier shall recover the candidate Anchor Event by traversing e references to a valid kind 1415 event carrying the same o value.

A.4 Anchor Event minimum shape

A kind 1415 Anchor Event shall contain:

{
  "kind": 1415,
  "tags": [
    ["o", "<artifact_sha256_hex>"],
    ["action", "issue"]
  ],
  "content": "<optional human-readable summary>"
}

It may include named structured metadata tags such as name, digest_generated_at, size_bytes, record_reference, record_description, domain, document_type, schema, or schema_digest.

A.5 Evidence Event minimum shapes

Every kind 1416 Evidence Event shall contain o, e, and action.

Additional minimum requirements are:

Action Additional required tag Meaning
initiate p intended transferee
accept none acceptance of referenced initiation
terminate none claimed lifecycle termination
attest none assertion about the record referenced by e
notice notice_type Anchor Publisher position concerning the artifact
encumber p beneficiary or secured party
discharge enc encumbrance event being discharged
redeem p obligor to whom performance is presented

Action-specific optional tags are defined in the OpenETR Nostr Wire Format Specification.

A.6 Content

Required machine-readable semantics shall be carried in event fields and tags. The content field should contain only a human-readable summary, comment, or other unstructured event data.

A.7 Query

A Nostr Verifier shall be capable of:

a) querying kind 1415 events using #o;

b) querying kind 1416 events using #o;

c) obtaining directly referenced events by event identifier where needed;

d) validating each event identifier and signature; and

e) reconstructing candidate graph paths using e references.

Relay OK responses and storage acceptance shall not be treated as semantic recognition of an event.

Annex B (informative) - OpenETR axioms and maxims

B.1 Ten axioms

  1. The digest identifies the artifact. A Digital Artifact is identified by a cryptographic digest, independently of its filename, location, format, or number of copies.
  2. A signature attributes a statement. Every OpenETR event is an immutable, attributable statement by a signing key. A valid signature establishes authorship and integrity; it does not, by itself, establish authority, recognition, or legal effect.
  3. An Anchor begins a candidate record. An Anchor Record establishes the starting point of a candidate Evidence Graph. It does not, by itself, establish uniqueness, validity, or recognition.
  4. Links construct the Evidence Graph. Evidence Graph records reference prior records and related evidence through cryptographic identifiers. Links establish claimed relationships; they do not, by themselves, establish validity or state.
  5. Events are evidence; state is derived. Events are not overwritten to represent current state. Consequential State is derived by evaluating signed Evidence Graph records according to defined rules.
  6. Invalid claims remain visible. Conflicting, unauthorized, or malformed statements remain part of the available evidence. A verifier warns about or excludes them according to its rule book rather than erasing signed history.
  7. A Digital Original has Consequential State. A Digital Original is a Digital Artifact for which Consequential State has been established through an Evidence Graph. Identical copies represent the same artifact; copying its bytes does not independently create another consequential history.
  8. Verification is separate from recognition. OpenETR verifies digests, signatures, record links, graph structure, and protocol transitions. External systems, policies, institutions, and laws determine recognition and effect.
  9. Identity is actor-neutral and contextual. A KBI identifies public-key verification material used to attribute signed evidence. Whether it is associated with a person, organization, service, device, or autonomous agent, and whether that actor is recognized or authorized, is determined by the applicable context.
  10. Evidence Graph records are portable across systems and domains. Signed records do not belong to one application, relay, operator, or jurisdiction. Domain adapters translate business actions into the general OpenETR model, while any conforming system can store, retrieve, and verify the resulting Evidence Graph.

B.2 Five maxims

  1. Digests identify. A cryptographic digest establishes which exact Digital Artifact the evidence concerns, independently of its filename, location, or number of identical copies.
  2. Signatures attribute. A valid signature establishes that a particular signing key made a statement concerning the artifact, without by itself proving the signer's identity, authority, or recognition.
  3. Links order. Cryptographic references connect signed records into an Evidence Graph and establish their claimed relationships, without by themselves deciding whether those records are valid or effective.
  4. Rules determine what follows. Defined protocol and verifier rules evaluate the available Evidence Graph records and derive the Consequential State that the evidence supports.
  5. Recognition gives effect. A relying party, institution, agreement, or law determines whether to accept the evidence or derived state for a stated purpose and what consequence that acceptance produces.

The first three maxims describe portable cryptographic evidence. The fourth describes the derivation of Consequential State. The fifth preserves the boundary between what OpenETR can verify and the effect that others choose or are required to give the result.

Annex C (informative) - Integration model

C.1 Host-system boundary

An integrating system commonly remains responsible for:

  • user accounts and authentication;
  • session and authorization management;
  • key custody and recovery;
  • document storage and movement;
  • domain validation and workflow;
  • KYC, AML, licensing, and registry checks;
  • privacy, disclosure, and retention; and
  • legal and operational recognition.

OpenETR commonly supplies:

  • artifact identification;
  • signed record construction;
  • publication and retrieval;
  • graph traversal;
  • structural and cryptographic verification;
  • verifier-policy findings; and
  • Consequential State derivation.

C.2 Integration patterns

Typical patterns include:

a) an account-based system hiding key and relay mechanics from users;

b) a host application importing the OpenETR Python component;

c) an automation or agent invoking the CLI with JSON output;

d) a service exposing REST APIs;

e) independent implementations exchanging Nostr-profile events; and

f) local archival and verification without a network dependency.

Annex D (informative) - Domain examples

D.1 Warehouse receipts

A warehouse receipt adapter can map:

Domain concept OpenETR concept
issued warehouse receipt file Digital Artifact
warehouse issuance record Anchor Record
warehouse operator signing identity Commitment Profile
selected facility or issuer Acting Profile
holder transfer initiate and accept records
pledge or lien encumber record
release discharge record
presentation for delivery redeem record
completed lifecycle terminate record

The adapter and applicable law determine whether this evidence satisfies MLWR, MLETR, registry, licensing, protected-holder, priority, or delivery requirements.

D.2 Other domains

The same core may support bills of lading, product passports, credentials, permits, health records, Apostille-related evidence, and other records. Each domain profile should define its own artifact boundary, action mappings, structured data, verifier policy, recognition inputs, and legal limitations.

Annex E (informative) - Suggested conformance statement

implementation: Example OpenETR Verifier
standard_draft: OPENETR/CD 1:2026
conformance_targets:
  - Core Verifier
  - Nostr Verifier
wire_binding:
  name: OpenETR Nostr
  kinds: [1415, 1416]
digest_algorithm: SHA-256
signature_algorithm: secp256k1 Schnorr
actions:
  - issue
  - initiate
  - accept
  - terminate
  - attest
  - encumber
  - discharge
  - redeem
verifier_policy:
  id: org.example.openetr.baseline
  version: 1.0.0
domain_profiles:
  - org.example.mlwr
deviations: []

Bibliography

  • OpenETR, Evidence Graph Design Note.
  • OpenETR, OpenETR Generic Verifier Policy.
  • OpenETR, OpenETR Generic Domain Adapter Specification.
  • OpenETR, Root and Profile Identity Model.
  • OpenETR, OpenETR Actor-Neutral Identity Design Note.
  • OpenETR, System Integration Considerations.
  • OpenETR, OpenETR CLI JSON Model.
  • OpenETR, Event Kind Registry.
  • OpenETR, Linked Evidence Record Kind Design Note.
  • OpenETR, Agentic AI Needs Consequential Evidence.
  • United Nations Commission on International Trade Law, Model Law on Electronic Transferable Records.
  • United Nations Commission on International Trade Law and UNIDROIT, Model Law on Warehouse Receipts.