NEXUS TRUST FRAMEWORK

Evidence, Provenance and Trust Classification

NEXUS Trust Framework defines a structured method for describing the evidentiary strength, provenance, review state, reproducibility and independent use associated with records maintained within the NEXUS infrastructure.

NEXUS does not attempt to determine whether an organization, person, standard or technology is universally "trusted".

Instead, NEXUS describes:

WHAT IS KNOWN

WHERE IT CAME FROM

HOW IT WAS VERIFIED

WHO REVIEWED IT

WHETHER IT CAN BE REPRODUCED

AND WHAT INDEPENDENT EVIDENCE EXISTS

Trust in NEXUS is therefore treated as an evidence problem rather than a declaration.


1. PURPOSE

Technical ecosystems contain many different types of claims.

A publisher may make a claim.

An implementation may make a claim.

A researcher may publish an observation.

A reviewer may verify a document.

An independent evaluator may reproduce a test.

A community may adopt a technology.

These events do not have identical evidentiary value.

NEXUS Trust Framework provides a common classification layer for representing these differences.

The framework is intended to make technical trust:


2. CORE PRINCIPLE

NEXUS does not assign trust merely because a record exists.

The central principle is:

TRUST SHOULD BE TRACEABLE TO EVIDENCE.

A NEXUS trust classification therefore SHOULD be connected to:


3. TRUST IS CONTEXTUAL

NEXUS does not define one universal trust score for every situation.

Trust depends on the object and the claim being evaluated.

For example:

A document may be authentic but technically incorrect.

A specification may be authoritative but not implemented.

An implementation may be widely used but not formally conformant.

A test may be reproducible but limited to one version.

A claim may be independently reviewed but later become obsolete.

Therefore, NEXUS trust records SHOULD identify the specific claim, object, scope and evidence under consideration.


4. TRUST RECORD

A NEXUS Trust Record SHOULD contain:


5. TRUST CLASSES

NEXUS defines the following internal trust classes.

T0 — UNVERIFIED

The record exists, but no meaningful verification of the relevant claim has been established.

Characteristics may include:

T0 does not mean that the claim is false.

It means that NEXUS has insufficient evidence to assign a stronger classification.


6. T1 — SOURCE VERIFIED

The primary source or originating document has been identified and verified to the extent reasonably possible.

Typical evidence may include:

T1 establishes source provenance.

It does not establish technical correctness, conformance or universal acceptance.


7. T2 — INDEPENDENTLY REVIEWED

The relevant record or evidence has undergone independent review.

The reviewer is separate from the original claimant or publisher for the purpose of the relevant evaluation.

Review may examine:

T2 does not automatically mean that the reviewed technical object is correct in every respect.

It means that an independent review process has occurred within the declared scope.


8. T3 — IMPLEMENTED

The referenced specification, protocol, model or technical concept has been demonstrated in an identifiable implementation.

Evidence may include:

T3 establishes implementation evidence.

It does not automatically establish conformance.

An implementation can exist without satisfying every requirement of a specification.


9. T4 — INTEROPERABILITY TESTED

The implementation or technical object has been evaluated through a defined interoperability procedure.

Evidence may include:

T4 SHOULD identify:

Interoperability is always bounded by the tested conditions.

Successful interoperability testing does not establish unlimited compatibility with all implementations or future versions.


10. T5 — INDEPENDENTLY ADOPTED

The technical object has evidence of independent adoption beyond its originating claimant or development environment.

Possible evidence may include:

T5 represents strong evidence of independent existence or use.

It does not mean that NEXUS declares the object universally superior, correct, safe or authoritative.


11. TRUST CLASSIFICATION IS NOT A QUALITY RANKING

T0–T5 are evidence classifications.

They are not a simple:

bad → good

ranking.

For example:

A highly authoritative document that has never been independently implemented may have strong source provenance but limited implementation evidence.

A widely implemented community technology may have extensive adoption evidence while lacking formal standardization.

NEXUS therefore treats trust as multidimensional.


12. TRUST DIMENSIONS

NEXUS may record separate dimensions.

SOURCE

How clearly can the origin of the information be established?

PROVENANCE

Can the history of the record be traced?

EVIDENCE

What evidence supports the claim?

REVIEW

Has an independent party examined the relevant information?

REPRODUCIBILITY

Can the result be independently reproduced?

IMPLEMENTATION

Does an identifiable implementation exist?

INTEROPERABILITY

Has interaction between implementations been demonstrated?

ADOPTION

Is there evidence of independent use?

RECENCY

How current is the evidence?

SCOPE

How broad or narrow is the claim?


13. PROVENANCE

Every important NEXUS assertion SHOULD have provenance whenever technically possible.

Provenance may identify:

Provenance allows users and machines to reconstruct the origin of an assertion.


14. PRIMARY SOURCES

NEXUS prefers primary sources when available.

Examples include:

Secondary sources may assist discovery.

However, a secondary source SHOULD NOT silently be represented as the original authority.


15. SOURCE AUTHENTICITY

NEXUS may distinguish:

CANONICAL

The source is identified as the authoritative publication location by the originating ecosystem.

VERIFIED

The source has been independently checked against available evidence.

SECONDARY

The source reproduces, summarizes or references information originating elsewhere.

UNKNOWN

The provenance cannot currently be established with sufficient confidence.


16. EVIDENCE CLASSES

Evidence may be classified as:

E0 — NONE

No supporting evidence is available.

E1 — ASSERTION

A claim exists but is primarily supported by a declaration.

E2 — DOCUMENTARY

Documentary evidence supports the claim.

E3 — REPRODUCIBLE

The relevant observation or test can be independently reproduced.

E4 — INDEPENDENT

Evidence originates from an independent party or environment.

E5 — MULTI-SOURCE

Multiple independent evidence sources converge on the same claim.

These classes may be used alongside T0–T5.

They do not replace the trust classification.


17. REVIEW STATES

NEXUS review status may be:

NOT-REVIEWED

No independent review has been recorded.

IN-REVIEW

A review is currently being conducted.

REVIEWED

The relevant material has been examined.

VERIFIED

The specific claim has been checked against identified evidence.

DISPUTED

A credible challenge has been recorded.

WITHDRAWN

The previous review status is no longer considered active.


18. INDEPENDENCE

Independence is evaluated relative to the claim being reviewed.

A reviewer should not be treated as independent merely because a different name appears on a document.

Potential conflicts may include:

Material conflicts SHOULD be disclosed.


19. CONFLICT OF INTEREST

NEXUS maintainers and reviewers SHOULD disclose material conflicts of interest.

A conflict does not automatically invalidate a review.

However, undisclosed material conflicts may reduce the evidentiary value of the review.

NEXUS records SHOULD distinguish:

WHO MADE THE CLAIM

from

WHO REVIEWED THE CLAIM

from

WHO CONTROLS THE EVIDENCE


20. REPRODUCIBILITY

A claim becomes stronger when an independent party can reproduce the relevant observation.

A reproducibility record may contain:

Reproducibility is particularly important for:


21. TEMPORAL VALIDITY

Trust is not necessarily permanent.

A record may become obsolete because:

NEXUS SHOULD therefore record timestamps and versions whenever possible.


22. TRUST DECAY

NEXUS may mark evidence as requiring reassessment when it becomes sufficiently old or when material circumstances change.

Trust decay does not mean:

old = false

It means:

old evidence may require renewed evaluation.

The appropriate reassessment interval depends on the type of object.


23. DISPUTED INFORMATION

A disputed record SHOULD NOT automatically be treated as false.

A dispute should identify:

NEXUS may preserve both the original assertion and the documented challenge.

This creates an auditable history.


24. CORRECTION

NEXUS records SHOULD be corrected when reliable evidence demonstrates:

Corrections SHOULD preserve historical information when technically and legally feasible.


25. TRUST AND CONFORMANCE

Trust classification and conformance classification are separate.

A record can have:

high evidentiary provenance

without being:

conformant

Similarly, an implementation may:

claim conformance

without having:

independent evidence

The NEXUS model therefore keeps:

TRUST

and

CONFORMANCE

as separate but connected layers.


26. TRUST AND AUTHORITY

Trust classification does not create institutional authority.

A T5 NEXUS record does not mean:

It only means that the available evidence satisfies the NEXUS criteria for the relevant classification within the declared scope.


27. TRUST AND STANDARDS

NEXUS may describe standards published by:

The original publisher remains authoritative for the official status of its own publication.

NEXUS provides an independent reference layer.


28. TRUST GRAPH

Trust information can be represented as a graph.

Example:

CLAIM

↓

SOURCE

↓

EVIDENCE

↓

REVIEW

↓

REPRODUCTION

↓

IMPLEMENTATION

↓

INTEROPERABILITY

↓

ADOPTION

This allows machines to distinguish between a declaration and a progressively stronger evidence chain.


29. MACHINE-READABLE TRUST RECORD

A NEXUS Trust Record MAY be represented as:

{

  "id": "NS:TRUST:000001",

  "subject": "NS:NEXUS:EXAMPLE:1.0",

  "classification": "T3",

  "evidenceClass": "E3",

  "sourceStatus": "VERIFIED",

  "reviewStatus": "REVIEWED",

  "reproducibility": true,

  "implementationEvidence": true,

  "interoperabilityEvidence": false,

  "independentAdoptionEvidence": false,

  "scope": "declared-scope",

  "evaluationDate": "2026-09-01",

  "registry": "https://www.null-state.dev/trust"

}


This record is a NEXUS classification.

It does not represent certification or accreditation issued by another organization.


30. TRUST SCORE

NEXUS SHOULD NOT reduce all trust dimensions to a single numerical score by default.

A single number can conceal important differences.

For example:

Source = VERIFIED

Evidence = REPRODUCIBLE

Review = INDEPENDENT

Adoption = UNKNOWN

is more informative than:

Trust = 82/100

The NEXUS model therefore prefers structured evidence dimensions over opaque aggregate scores.


31. CLAIM TYPES

Trust evaluation may apply to different claim types.

IDENTITY CLAIM

What is this object?

PROVENANCE CLAIM

Where did this object originate?

RELATIONSHIP CLAIM

How is this object related to another object?

IMPLEMENTATION CLAIM

Does an implementation exist?

CONFORMANCE CLAIM

Does an implementation satisfy specified requirements?

INTEROPERABILITY CLAIM

Can two systems interact under defined conditions?

ADOPTION CLAIM

Is the technology independently used?

STATUS CLAIM

What status does the originating publisher assign to the object?

Each claim type may require different evidence.


32. STATUS CLAIMS

NEXUS must distinguish:

ORIGINAL STATUS

from

NEXUS CLASSIFICATION

For example:

Original publisher status: Recommendation

may coexist with:

NEXUS trust classification: T1

These statements describe different things.

NEXUS must never silently replace the publisher's terminology with its own.


33. TRUST TRANSITIONS

A record may evolve over time.

Example:

T0 → T1 → T2 → T3 → T4 → T5

This sequence is possible but not mandatory.

A record may move between classes depending on evidence.

A T5 classification today does not guarantee T5 status indefinitely.

Trust classifications SHOULD therefore be timestamped.


34. EVIDENCE FAILURE

If evidence supporting a classification becomes invalid, unavailable or disproven, NEXUS may lower or suspend the relevant classification.

Possible states include:

The historical record SHOULD remain available where appropriate.


35. TRUST INTEGRITY PRINCIPLES

NEXUS Trust Framework follows these principles:

1. PROVENANCE

Claims should be traceable to identifiable sources.

2. EVIDENCE

Claims should be distinguished from supporting evidence.

3. INDEPENDENCE

Independent review should be distinguishable from self-declaration.

4. REPRODUCIBILITY

Reproducible observations should be distinguished from non-reproducible claims.

5. SCOPE

Trust classifications apply within declared scopes.

6. TEMPORALITY

Evidence may change over time.

7. TRANSPARENCY

Material limitations should be visible.

8. CORRECTABILITY

Demonstrated errors should be correctable.

9. DISPUTABILITY

Credible challenges should be recordable.

10. NON-DECEPTION

NEXUS classifications must never be represented as authority belonging to another organization.


36. RELATION TO THE NEXUS SYSTEM

The NEXUS Registry answers:

WHAT EXISTS?

The NEXUS Knowledge Graph answers:

HOW IS IT CONNECTED?

The NEXUS Conformance Framework answers:

DOES IT SATISFY THE REQUIREMENTS?

The NEXUS Trust Framework answers:

HOW STRONG IS THE AVAILABLE EVIDENCE?

The Evidence & Provenance Layer answers:

WHERE DID THE EVIDENCE COME FROM?

The Governance Constitution answers:

HOW ARE THESE SYSTEMS GOVERNED?

Together these components form the NEXUS reference infrastructure.


37. DESIGN PRINCIPLE

NEXUS does not ask users to trust NEXUS merely because NEXUS says something is trustworthy.

Instead, NEXUS attempts to expose the evidence behind the classification.

The preferred relationship is:

CLAIM

→ SOURCE

→ EVIDENCE

→ REVIEW

→ REPRODUCTION

→ INDEPENDENT USE

The more transparent this chain becomes, the more useful the record becomes to both humans and machines.


38. FINAL PRINCIPLE

The NEXUS Trust Framework is based on a simple rule:

DO NOT HIDE THE EVIDENCE CHAIN.

A trustworthy technical registry should allow an observer to move backwards from a conclusion toward its supporting evidence.

A machine should be able to determine:

what is claimed,

who claims it,

where the information originated,

what evidence exists,

whether the evidence was independently reviewed,

whether the result can be reproduced,

whether an implementation exists,

whether interoperability was demonstrated,

whether independent adoption exists,

and whether the information remains current.


NEXUS TRUST FRAMEWORK

CLAIM

↓

SOURCE

↓

PROVENANCE

↓

EVIDENCE

↓

REVIEW

↓

REPRODUCIBILITY

↓

IMPLEMENTATION

↓

INTEROPERABILITY

↓

INDEPENDENT ADOPTION


Scope Statement

NEXUS Trust Framework is an independent technical evidence-classification framework.

Its classifications describe the evidence available to NEXUS for a defined object, claim and scope.

They do not constitute governmental approval, regulatory approval, accreditation, certification, legal recognition or endorsement by another standards organization.

NEXUS does not ask the world to trust a declaration.

NEXUS makes the evidence chain inspectable.