About Loreth: Identity Across Change

Institutional Foundation Series · Inaugural publication

An introduction to the problem of keeping intelligent agents identifiable, accountable and governable as the circumstances around them change.

White Book · Edition 1.0 · Approved for Production · 16 October 2026

Published by the Loreth Registry Foundation Founded by J. Christopher Holmes

Publication record

Title: About Loreth: Identity Across Change Series: Institutional Foundation Series Publication type: White Book Edition: 1.0 Permanent publication identifier: LRF-WB-001 Publication status: Approved for Production Official publication date: 16 October 2026 Owner and publisher: Loreth Registry Foundation Supersession history: Supersedes no prior publication; has not yet been superseded by a later edition

Published by the Loreth Registry Foundation Founded by J. Christopher Holmes

Purpose and scope

This publication introduces Loreth’s institutional purpose. It explains the governance problem that precedes technical specifications: how to keep an intelligent agent’s identity, accountable custody and history understandable across change.

It draws on the current Registry Handbook and distinguishes governing principles from proposed Registry arrangements. It does not establish certification, legal recognition, consensus adoption or a guarantee about an agent’s behavior. The technical proposal described in the Handbook remains a proposal. The website’s separately described Public Registry Environment—Reference Edition does not change that status. [1] [2]

Copyright and interpretation

Copyright © 2026 Loreth Registry Foundation. All rights reserved. Rights language is subject to confirmation before approved release.

Published by the Loreth Registry Foundation Founded by J. Christopher Holmes

The four foundational principles are design commitments. Descriptions of the Registry concern the proposed framework unless another status is explicitly stated. The procurement-agent narrative is a constructed illustration, not a reported deployment or incident. External references provide context, not endorsement of Loreth. Future questions are identified as unresolved.

Registration is not endorsement. Verification is not certification. An authentic record does not determine whether an agent is safe, lawful, accurate, suitable or authorized for a particular action. [1]

Contents

• Executive summary
• 1. Identity is familiar. Continuity is less visible.
• 2. An agent changes. What, exactly, continues?
• 3. The Agent Identity Gap
• 4. Identity Across Change
• 5. Why Loreth exists
• 6. Four foundational commitments
• 7. Trust needs evidence—and boundaries
• 8. An institution must remain accountable too
• 9. A public body of work
• 10. The beginning of continued exploration
• Glossary
• References and source notes
• About the publisher
• Colophon

Executive summary

An intelligent agent can acquire a new name, model, key, deployment environment or accountable party without becoming unrelated to its past. Yet the connection between its earlier and current state may be difficult to establish. The software may keep working while the basis for accountability becomes unclear.

Loreth calls this the Agent Identity Gap. The issue is not the absence of all identity technology. Authentication, authorization, monitoring and organizational records already perform essential work. The question is how their evidence remains connected to an identifiable subject through legitimate change. [1]

Identity Across Change is Loreth’s governing idea. The Loreth Registry Foundation develops public-interest registry and governance infrastructure around four commitments: Identity Persists, Custody Remains Accountable, Continuity Is Planned and Integrity Is Verifiable. The proposed Registry is intended to make defined claims about identity, custody, lifecycle and record history reviewable over time. [1] [2]

These claims have boundaries. Registration does not endorse an agent. Custody does not settle legal ownership. Continuity does not mean that an agent must operate forever. Verification does not guarantee the truth of every assertion or the safety of future behavior.

This publication begins with the problem rather than the institution. Its purpose is to make the question understandable enough to examine: when an agent changes, what must remain identifiable—and who must remain accountable?

1. Identity is familiar. Continuity is less visible.

We already live among systems that make subjects recognizable. People have identity records. Organizations have registrations. Devices, services, domains and assets have identifiers appropriate to their settings. These arrangements differ, but they share a practical purpose: helping people and systems distinguish one subject from another. [1]

An identifier becomes useful when it connects to something more than a label. A record can help us understand the subject’s status, find the relevant responsible party or trace a change. The familiar question “Which one?” begins to support another question: “What happened to it?”

The Legal Entity Identifier offers an institutional example. It connects a unique identifier to verified reference information about a legal entity. Its purpose is identification, supported by maintained data—not a general verdict on the entity’s conduct. This is an example of the wider value of governed reference systems, not a claim that Loreth shares the LEI’s legal, standards or institutional standing. [3]

The need for continuity is easiest to overlook when nothing is changing. A current name and a current record may appear sufficient. But an organization can rename itself. A device can receive a new credential. A service can move to another environment. When that happens, we need to understand whether we are looking at the same subject, a replacement or a different subject using a familiar name.

Identification alone does not resolve every question. It does not establish that an action was permitted or that a subject deserves trust. It gives those inquiries something definite to refer to.

That is the starting point for Loreth: important decisions require more than recognition at a single moment. They require a way to understand a subject across time.

Recognition begins with identity. Accountability needs a history.

2. An agent changes. What, exactly, continues?

Industry Example—illustrative. The following procurement agent is a constructed example used throughout this publication.

Consider an organization that operates an agent to assist with procurement. It gathers supplier information, prepares recommendations and performs tasks within permissions set by the organization. Its activity appears in operational records. People have approved its use for a defined purpose.

Then the organization moves the agent to another environment. The service accounts change. A new signing key is issued. The model is updated. A different team takes over maintenance.

The migration succeeds. The agent performs its tasks again.

But a later reviewer asks a question that successful deployment did not answer: is this the agent to which the earlier approval applied?

The reviewer may also need to know whether the earlier incident belongs to this agent’s history, who approved the change and who is now accountable for its record. The relevant facts could be distributed across deployment logs, tickets, account directories and recollections. None need be false for their relationship to be unclear.

The agent’s familiar name is not enough. A new credential may establish a current technical relationship without explaining the old one. The continued purpose may suggest continuity without demonstrating that the change was authorized and documented.

This is not an argument against migration, new models or new credentials. Those changes may be appropriate. It is an argument for preserving the basis on which people can understand them. The Handbook uses platform migration to illustrate precisely this distinction. [1]

Now imagine a more difficult transition. The responsible vendor exits. A custodian cannot be contacted. Two organizations disagree over control. An agent is retired, but a later investigation still needs to establish which approvals and actions belonged to it.

In each case, the question is less about whether software exists than whether responsibility can still be followed.

Working again is not the same as being accountable across change.

3. The Agent Identity Gap

The Agent Identity Gap is not an empty space where no identity systems exist. It is the difficulty of keeping identity, authority, responsibility and history connected when the surrounding technical and organizational facts change. [1]

A model name identifies a technology. It does not, by itself, distinguish one organization’s agent deployment from another. A workload credential participates in authentication. It does not, by itself, provide the agent’s complete operational history. A log can record an event. Its usefulness for a later inquiry depends partly on whether the subject of that event remains identifiable.

Authorization answers another question: what may this agent do here, under these conditions? An enduring identity does not create those permissions. A permitted action also does not establish the continuity of every record associated with the agent.

Existing arrangements can preserve history, connect credentials and govern change. Loreth does not claim that they are incapable of doing so. The concern is whether the connections remain clear across platforms, organizations and lifecycle events, especially where other parties need to inspect them. [1]

This is also part of a wider conversation about agents. NIST has described accountability gaps created by credential sharing and the need to treat agents as identifiable entities with their own credentials and entitlements. It also points to existing protocols that can address important identification and authorization needs. That work supplies context; it does not validate Loreth’s particular proposal. [4]

The distinction matters most where the agent has consequential authority, sensitive access, persistent relationships or a history on which others rely. Loreth’s framework does not argue that every temporary task needs public registration. The controls should be proportionate to the consequences of ambiguity. [1]

The gap appears when several locally understandable facts no longer add up to an understandable history.

A credential can change without the subject changing.

4. Identity Across Change

Loreth Principle—governing idea, not a claim that every continuity question has been resolved.

Identity Across Change does not mean that an agent’s components remain unchanged. It means that legitimate changes can be connected to an enduring subject through a governed record. [1] [5]

The distinction is important. If persistence required the same model, host, key and team forever, identity would break whenever responsible maintenance occurred. If persistence required no explanation at all, almost any replacement could claim the old history.

A durable approach must avoid both mistakes. It permits change while requiring the continuity claim to be intelligible.

For the procurement agent, a migration may become a documented event in the history of the same subject. Its current technical facts differ from its earlier ones. The relationship between them is maintained rather than inferred from a reused name.

Not every apparent continuation deserves that treatment. A copy, fork or independently operated reimplementation may require a separate identity. A disputed change may require uncertainty to remain visible. Similarity of code, purpose or name does not settle the matter. The Handbook identifies these equivalence questions as work requiring further examination. [1]

This is not a claim about an agent’s consciousness, personhood or resemblance to a human self. It is a practical governance question: what subject do these records concern, and why should a later reader understand the records as connected?

The answer requires evidence and attributable decisions. Persistence is not achieved by declaring that everything belongs to the same history. It is achieved by making the basis for that history reviewable.

What persists must be explained, not assumed.

5. Why Loreth exists

The Loreth Registry Foundation develops public-interest registry and governance infrastructure for durable agent identity, accountable custody, planned continuity and verifiable governance. Its interest is not artificial intelligence as a market category. Its interest is what happens to accountability when the systems and circumstances around an agent change. [1] [2]

Loreth’s proposed Registry provides an enduring reference for defined claims about an agent. The purpose is to make the subject identifiable, the current accountable custody clear, the lifecycle understandable and the record’s history inspectable. A Registry in this sense is governed reference infrastructure, not a directory of promotional profiles.

The distinction from other kinds of systems is deliberate. Loreth is not a model provider or an agent marketplace. It does not replace the controls that authenticate a workload, authorize an action, monitor behavior or determine whether a particular deployment is appropriate. It provides a subject reference that those controls and the people using them can follow across change. [1]

The institution’s concern begins before implementation detail. What should remain attributable after a migration? What should happen when custody is uncertain? What evidence should a later reviewer be able to inspect? Which parts of a record should be public, and which should remain protected?

These questions are relevant because reliance often extends beyond the team operating the agent. An auditor, partner, affected person or successor team may need to understand the history without having participated in the original decision.

A useful response must therefore do more than collect information. It must explain the status of the information, the authority behind changes and the limits of what the record supports.

That is the institutional purpose of Loreth’s proposed work.

The record is a reference for accountability, not a verdict on conduct.

6. Four foundational commitments

Loreth Principles—design commitments. They are introduced here conceptually; the Registry Handbook explains the proposed framework in greater detail.

The four principles name different obligations. Together, they ask whether a subject can remain identifiable, accountable and reviewable through change. None substitutes for the others. [1]

Identity Persists

Which subject continues through legitimate change?

An agent should not become disconnected from its relevant history merely because a key rotates, a deployment moves or a model changes. The enduring reference is meant to connect accepted changes, not freeze the technical system.

Return to the procurement agent. Its new service account should not silently erase the relevance of an earlier approval. Nor should the old identity be borrowed by an unrelated replacement simply because a familiar name is convenient.

The principle therefore has two sides: preserve continuity where it is supported, and distinguish subjects where it is not. The continuity relationship must be governed rather than guessed.

Identity should survive legitimate change.

Custody Remains Accountable

Who governs the active record now?

Someone must be accountable for maintaining the record, responding to relevant events and supporting changes with the required authority and evidence. In Loreth’s proposed framework, the current primary custodian serves that record-governance role. [1]

Custody is not a universal determination of ownership or liability. The developer, model provider, host and operator may be different parties. Their legal duties remain subject to law and other governing arrangements. A Registry record does not settle them all by naming one person or organization.

When the procurement agent moves to a new team, the new entry should mean more than “this name was typed into a field.” The transition needs an intelligible basis. The current accountable party should be clear without erasing the previous chain.

A name in a plan is not authority.

Continuity Is Planned

What route exists before disruption?

A continuity problem is easier to govern before urgency narrows the choices. The relevant questions include who may initiate a change, what evidence is needed, which restrictions apply and what happens when a proposed successor cannot or should not take over. [1]

Naming a successor does not make that party the custodian. A potential route and an accepted transfer are different things.

Continuity also includes responsible endings. An agent may need to be suspended or retired rather than transferred. Preserving its history can remain important after active operation ends. The purpose is not to ensure that every agent survives; it is to avoid abandoning governance when circumstances change.

For the procurement agent, that could mean a documented recovery route after key loss—or a decision that retirement is the appropriate outcome.

Continuity is not survival.

Integrity Is Verifiable

Can the accepted record and its history be checked?

A later reader should be able to inspect the record’s integrity, the relationship between versions and the evidence supporting relevant changes. The history should not depend entirely on the memory of the people who maintained it. [1]

The procurement agent’s migration should leave something more durable than the assertion “we handled that.” The record should support reconstruction of what changed and the basis on which it was accepted.

Verification has a scope. It can establish particular record properties without establishing that every underlying assertion is true. An authentic statement can be mistaken. A correctly preserved history can include a failure.

The principle is therefore a commitment to checkability, not an offer of infallibility.

A verifiable statement can still have limits.

7. Trust needs evidence—and boundaries

One of the easiest mistakes is to make identity carry claims it cannot support.

A registered agent is identifiable under the applicable Registry arrangements. That does not make the agent endorsed. A verifiable record can support defined assertions about its integrity and history. That does not certify the agent’s behavior. The distinction is part of responsible reliance, not a disclaimer detached from the framework. [1]

The same restraint applies to authorization. A record may be authentic while the agent has no permission to perform a particular transaction. Whether it may act depends on the applicable policy, law, contract and technical controls.

Risk and evidence must remain separate too. An agent operating in a consequential setting does not acquire stronger evidence because the consequences are greater. More serious use creates a stronger reason to examine the basis for reliance; it does not supply that basis. [1]

Authenticity is not authorization. More consequence is not more evidence.

Public accountability introduces another distinction. Reviewable claims do not require indiscriminate exposure. A public record should not publish private keys, confidential agent memory or unnecessary sensitive information. The question is which information lets others check the permitted claim while protecting what should not be public. [1]

Publicly checkable. Not publicly exposed.

These limits do not make the infrastructure less useful. They make its usefulness more precise. A relying party can ask what this record actually establishes, which information is current, what uncertainty remains and what further checks are needed.

Where the evidence is insufficient, uncertainty should stay visible. Where authority is disputed, the record should not turn a convenient resolution into an invented agreement.

A trustworthy institutional practice begins by refusing to make a record say more than its evidence supports.

8. An institution must remain accountable too

The argument for continuity applies to Loreth’s own work.

If publications can be revised without a visible history, if commitments change without an intelligible reason, or if uncertainty disappears from the public account, the institution would weaken the discipline it asks readers to consider. The Publication System therefore calls for permanent publication identifiers, preserved editions and transparent publication history. [5]

An enduring institution needs to distinguish what it is committing to from what it has demonstrated. It also needs to distinguish an established rule from a question still under examination. Institutional presentation cannot substitute for that work.

Loreth’s purpose connects several kinds of activity. Foundation stewardship concerns the framework and the durability of its body of work. The proposed Registry concerns identifiable subjects and governed records. Digital interfaces make records and publications accessible; they do not turn the institution’s purpose into software-product positioning. Community participation brings questions and scrutiny from people who encounter different consequences and constraints. [1] [2] [5]

This is not a claim that all future governance arrangements are settled or operating. It is an account of the work that must support the institutional purpose. Credible stewardship requires evidence of its own practices as they develop.

The community’s role is therefore more than following announcements. A useful participant can identify a weak distinction, test an example, question an unsupported claim or contribute experience from a difficult transition. Such contributions can make the framework more accountable to the settings in which it might be used.

Loreth’s ambition is not exhausted by publishing a standard. A standard belongs to a wider record of explanation, revision, examination and stewardship. That record must remain understandable to people who arrive after the original decisions were made.

Stewardship must outlast the circumstances in which it began.

9. A public body of work

Future Possibility—an agenda for examination, not a set of settled outcomes or announced operating programs.

The question of identity across change does not end with the four principles. It opens a body of work.

When is a copy a new subject? Which changes support continuity, and which require a different identity? How should disputed custody remain visible? What evidence is sufficient for a particular claim, and when does it become stale? How can public verification protect sensitive information? How can records remain understandable across independent operators? The Handbook identifies these as matters needing further work. [1]

The answers should be tested against real transitions: migration, recovery, organizational change, contested control and retirement. A framework that reads clearly in a document may encounter difficult cases in practice. Those cases belong in the work, not outside its public account.

This gives different readers meaningful ways to examine Loreth. An engineer can question what continuity requires when a deployment is replicated. A governance professional can examine whether custody and dispute routes are sufficiently clear. An affected person can ask what information supports a public claim and how an error would be addressed.

No discipline answers all these questions alone. Technical integrity cannot settle every authority dispute. A governance rule cannot compensate for missing evidence. Public visibility cannot justify exposing protected information.

The institutional task is to keep the relationships clear enough that each question can be examined on its own terms.

A durable body of work will contain principles, proposed specifications, explanations, examples and unresolved questions. Their distinctions should remain visible as the work grows. That is how curiosity can become informed contribution without being mistaken for endorsement or consensus.

Open questions should remain visible.

10. The beginning of continued exploration

Guidance—reading paths, not a request to rely on an agent or a Registry claim.

The question that began this publication was simple: when an agent changes, what continues?

It leads to a practical institutional concern. People need to know which subject a record describes, who is accountable now, what happened before and what the available evidence can support. Those questions remain relevant when the original environment, credential or responsible team no longer exists.

Loreth is developing its body of work around that concern. Its value must be examined through the clarity of its distinctions, the quality of its evidence and the durability of its stewardship—not inferred from the name of the institution or the appearance of this publication.

The next step need not be agreement. It can be a more precise question.

Continue exploring

Core Principles is the next reading for understanding the four commitments and the different ambiguities they address.

Frequently Asked Questions is the next destination for testing an assumption: whether registration means endorsement, whether custody means ownership, or whether continuity means indefinite operation.

Registry Handbook is the next reading for examining how the proposed framework describes records, lifecycle events, custody, verification and responsible reliance.

Readers who want to contribute can continue toward the community’s published participation routes when available. The most useful contribution may begin with a case the framework does not yet explain well.

The work ahead is not to prevent intelligent agents from changing. It is to make change governable without losing the ability to understand what came before.

Change need not make responsibility untraceable.

Across history, communities have created identity records, stewardship systems, accountability frameworks and governance structures for the same underlying reason: change is inevitable, but continuity cannot be left to memory or assumption. Loreth begins from that realization. Its purpose is not to prevent change, but to help ensure that change does not erase responsibility, accountability, continuity or the ability to understand what came before. This publication is the opening of that body of work.

Glossary

Accountability. The ability to identify responsibility and examine the basis for relevant actions or claims. In this publication, Registry accountability concerns the record and its governance; it is not a universal allocation of legal liability.

Agent. An intelligent or automated system that can pursue tasks, use tools, retain context or relationships, communicate or act under delegated authority. This is an operational description, not legal personhood. [1]

Agent Identity Gap. The difficulty of keeping an agent’s identity, authority, accountable custody and history intelligibly connected through technical and organizational change.

Continuity. Planned arrangements for changes such as transfer, recovery, suspension and retirement. It does not require indefinite operation. [1]

Credential. Technical material used in authentication or access arrangements. It is not interchangeable with an enduring subject identity.

Custodian. The current primary party accountable for governing an active Registry record in its deployment context. Custody does not settle every ownership or liability question. [1]

Identity Across Change. Loreth’s governing idea that legitimate changes should remain connected to an identifiable subject through a governed and reviewable history.

Integrity. The record properties and relationships that allow accepted information and its history to be checked. Integrity alone does not prove every underlying assertion true.

Registry. Governed infrastructure that makes authoritative identifiers, current assertions, lifecycle rules and history resolvable. Authority is bounded by the applicable Registry arrangements. [1]

Stewardship. The responsibility to maintain the framework and its body of work with attention to evidence, revision, public understanding and long-term continuity.

Verification. Checking defined properties, evidence, integrity and authority under applicable rules. It is not certification of an agent’s conduct. [1]

References and source notes

Institutional source material

[1] Loreth Registry Foundation, Registry Handbook. Public guidance edition 1.0, October 2026. Immediate source for the four principles, proposed Registry role, custody, continuity, privacy, limitations and unresolved development questions. The Handbook describes LORETH-STD-001 Edition 2.1 as proposed; this publication makes no separate conformance claim.

[2] Loreth Registry Foundation, Loreth Registry Handbook—Production Reading / Brand Edition. Manuscript edition 1.0; production polish 15 October 2026. Source for the current institutional description and the distinction between the proposed framework and the website’s separately presented Reference Edition. This publication does not independently audit that environment.

[5] Loreth Publication System. Supplied eight-page institutional statement. Source for Identity Across Change as the governing idea, claim discipline, reader promise, institutional character, publication history and preserved editions.

External context

[3] Global Legal Entity Identifier Foundation, The Legal Entity Identifier (LEI). Consulted 16 October 2026. Used only as a bounded example of governed identifiers linked to reference information. No affiliation or equivalent institutional standing is asserted.

[4] National Institute of Standards and Technology, Back to the Future: Why Agentic AI Needs a Strong Identity Foundation. 27 August 2026; consulted 16 October 2026. Used for context on agent identity, authorization and accountability gaps from credential sharing. It does not endorse Loreth or demonstrate Loreth’s design.

Interpretation and draft provenance

The procurement-agent narrative is constructed to explain the concepts. It is not empirical evidence, a customer reference or a reported incident. The final sections distinguish institutional commitments, guidance and future possibilities from established operating arrangements.

This approved-for-production publication uses the Loreth Registry Handbook and Loreth Publication System as its principal institutional sources. The White Paper, Manifesto, About Loreth, Visual Direction and relevant LEOS materials remain part of the source family for controlled future revision. No LEOS-specific claim is introduced here. All references below are release-ready public references; no editorial workspace paths are retained.

[1]: https://loreth.org/registry-handbook "Loreth Registry Handbook, public guidance edition 1.0, October 2026" [2]: https://loreth.org/about-loreth "Loreth Registry Foundation, About Loreth publication family" [3]: https://www.gleif.org/en/about-lei/introducing-the-legal-entity-identifier-lei "GLEIF, The Legal Entity Identifier (LEI)" [4]: https://www.nist.gov/blogs/cybersecurity-insights/back-future-why-agentic-ai-needs-strong-identity-foundation "NIST, Back to the Future: Why Agentic AI Needs a Strong Identity Foundation, 27 August 2026" [5]: https://loreth.org/publication-system "Loreth Publication System"

About the publisher

The Loreth Registry Foundation develops public-interest registry and governance infrastructure for durable agent identity, accountable custody, planned continuity and verifiable governance. This publication opens the Institutional Foundation Series by explaining the problem and the commitments that give that work its purpose. It does not replace the Registry Handbook or a technical specification.

Published by the Loreth Registry Foundation Founded by J. Christopher Holmes

Colophon

About Loreth: Identity Across Change is the inaugural White Book of the Institutional Foundation Series. Publication identifier LRF-WB-001. Edition 1.0. Approved for Production. Official publication date: 16 October 2026.

Published by the Loreth Registry Foundation Founded by J. Christopher Holmes

The manuscript separates a conceptual introduction from implementation detail. Its recurring procurement-agent example is illustrative. Its chapter takeaways are intended to make distinctions recoverable after reading. Contents, glossary and references support scanning, skimming, study and return visits.

The companion production blueprint specifies a light-first reading edition using Manrope headings, Inter body text, the official Loreth mark and governed educational graphics. Final typesetting, accessibility review and press preparation remain production activities under the approved record; they must preserve the controlled text, metadata and edition history.

Publication-system purpose: making identity, accountability, continuity, trust, governance and stewardship understandable, memorable and worthy of continued exploration.