CHAPTER 03

Hybrid-ID: Identity, Authority, and Policy

How Hybrid-ID connects familiar sign-in to application permissions, optional trust, and the direction of accountable agent participation.

10 min read · Q3 2026
Hybrid-ID sign-in01App permissions02Permitted context03Action authority04
  1. Hybrid-ID sign-in
  2. App permissions
  3. Permitted context
  4. Action authority
Hybrid-ID sign-in connects to application permissions, permitted context, and a separately authorized action; each step retains its own authority check. Conceptual illustration.

Identity answers who is interacting with a system. Authority answers what that participant may do. Policy adds the conditions under which an otherwise permitted action should proceed. Hybrid-Chain treats these as related but distinct concerns, allowing an organization to describe responsibilities more precisely than a single all-purpose account role.

From identity to a bounded action

A user may belong to an organization, work within a particular workspace, and access only certain resources. An application or agent can require a narrower set of capabilities than its operator. Sensitive actions may also depend on the target resource, current lifecycle state, and an applicable review or approval. Integration design should preserve this context at the point of use rather than assuming that an earlier successful authentication authorizes every later request.

This supports separation of responsibilities. The person proposing an action, the reviewer evaluating it, the operator maintaining infrastructure, and the party authorized to execute it need not be the same participant. The appropriate arrangement depends on the product and deployment; the architecture does not equate infrastructure administration with unrestricted control over customer activity.

Decisions that remain attached to their subject

Governance is useful when a decision can be understood in relation to the proposal that was reviewed. A change of terms or version can change what must be evaluated. The platform's public governance model therefore emphasizes reviewable proposals and their eligible decisions, with execution treated as a separate responsibility. Applications should make that distinction visible rather than interpreting a favorable review as an automatic side effect.

Trust and compliance-oriented modules add another kind of boundary: a verifier may need to inspect the issuer, scope, or status of a credential without receiving the protected case material behind it. This makes verification useful without turning each check into a new distribution of confidential records. A credential's meaning still depends on its issuer and purpose; a technically valid record does not establish every possible claim about its subject.

Engineering references

These reads show how a governance review stays attached to an identifiable subject and its recorded decisions.

Find a proposal

Identify the governance subject that belongs to the review.

GET/api/v2/governance/proposals
Observable result
The proposal records visible to the authorized caller.
Authority and limits
Listing a proposal does not decide or execute it.

Inspect the reviewed subject

Connect a decision to a specific proposal rather than a display name.

GET/api/v2/governance/proposals/{proposal_uuid}
Observable result
Proposal state, authority, approvals, and retained evidence.
Authority and limits
Recorded approval is distinct from execution; confirm the subject and current state.

Selected operations were checked against the public OpenAPI contract on October 1, 2026. Links open documentation; they do not invoke an operation. Authentication, exact schemas, and deployment requirements remain defined by the current contract.

Assign authority before connecting the workflow.

A useful trust model names who can see information, change policy, approve a proposal, and execute an action. The roles below are a deployment-design worksheet, not a claim that a particular custody or key-sharing arrangement applies to every customer. One organization may perform several roles, but each responsibility should still have an explicit owner.

  1. Information owner
  2. Scoped participant
  3. Authorized decision
  4. Independent review

Information owner: decide what may be used

The customer assigns an owner to supplier records, internal instructions, and other protected material. That owner defines intended recipients, classification, and retention obligations. A collection publisher separately reviews what may become agent context; permission to inspect an object is not permission to distribute its contents.

Retain for review
A source owner, permitted audience, approved context reference, and handling policy.
Acceptance check
Can the agent perform its task using approved context without receiving the confidential source file?

Application or agent: act inside a delegated scope

The integrating organization owns the application's behavior and credential handling. The workload uses the exact permissions and request-signing requirements of its interfaces. A context-read credential does not confer wallet or funding authority. Keep signing material outside prompts, generated explanations, browser logs, and public evidence.

Retain for review
A workload identity, named credential custodian, allowed operations, and revocation procedure.
Acceptance check
Does removing one capability stop that action without requiring an all-powerful shared credential?

Policy owner and reviewer: bind a decision to its subject

The organization appoints the people who set spending conditions and review business proposals. A decision needs a specific subject: destination, network, asset, amount, and applicable policy context. If those terms change, the integration should obtain a fresh evaluation and the approvals required by the selected execution workflow.

Retain for review
The reviewed proposal, policy reference, decision reasons, and accountable reviewer.
Acceptance check
Can a reviewer tell whether the action presented for execution is the action that was actually reviewed?

Execution and key custodians: authorize the financial step

The customer and provider must agree who holds key material or signing shares, who participates in authorization, and who controls recovery for the chosen wallet arrangement. Infrastructure operation, application administration, and asset authority are different responsibilities. This public architecture does not prescribe or disclose a proprietary signing configuration.

Retain for review
A deployment-specific authority matrix, key/share custody allocation, and tested recovery ownership.
Acceptance check
Which participants can authorize, sign, halt, or recover the exact operation—and which cannot?

Service operator and external provider: own their respective stages

Each authoritative service enforces its documented resource boundary. Operators maintain the agreed service environment; external networks or payment providers apply their own acceptance and completion rules. The integration owner reconciles those stages instead of treating an application's success message as universal completion.

Retain for review
Named service contacts, escalation paths, network/provider references, and completion criteria.
Acceptance check
Who investigates an ambiguous outcome, and what authoritative record can resolve it?

Reviewer or auditor: check a scoped claim

A reviewer receives the evidence needed for a defined question, with confidential material disclosed only through an authorized process. Verification authority is not execution authority. A public reader may inspect a published record without being entitled to the private business context behind it.

Retain for review
A review purpose, permitted evidence set, verification procedure, and recorded conclusion.
Acceptance check
Does the conclusion stay within what the inspected records actually establish?

The result is a responsibility map that survives changes in software or personnel. Before deployment, name the actual organizations and people behind these roles, document unresolved boundaries, and test the intended separation of duties.

Delegation for applications and agents

Delegation makes automation practical when it can be limited to the intended task. An agent might be permitted to inspect a budget, retrieve an approved collection, or propose an action while remaining unable to execute a transfer. Permissions, resource boundaries, and decision history provide the operating context in which that agent can be useful and accountable.

The resulting advantage is controlled participation. Organizations can involve more applications, partners, and automation without requiring every participant to receive the same access. The integration must still establish the exact permissions and approval requirements of its chosen workflow.

Hybrid-ID: a familiar identity across applications

Hybrid-ID is the public identity experience built on Hybrid-Chain's identity infrastructure. It gives an established Hybrid account a familiar way into participating business applications, beginning with Hybrid-Chain itself. Its contribution is continuity of recognized identity: an application can associate a returning participant with the workspace and preferences it maintains, while applying its own membership, resource, and action permissions.

OpenID Connect sign-in is available for registered web applications. The public developer guide describes a server-based integration using Authorization Code with PKCE S256, application consent, exact registered HTTPS callback URLs, and approved scopes. The provider's live discovery document identifies auth.hybrid-id.com as the canonical issuer. Applications should use the maintained integration guide and discovery metadata rather than treating an account-password endpoint as a general-purpose SSO interface.

The application recognizes the identity through the validated issuer and subject, not an email address used as a universal account key. The provider supports pairwise subjects: unrelated application sectors receive different subject identifiers. That supports scoped recognition across integrations; it does not imply anonymity or authorize an application to correlate every account elsewhere. Each application retains its own session and application data.

Optional verification and privacy with a defined purpose

Basic authentication establishes an account interaction. A business can separately require an appropriate credential or verification before enabling a particular feature. The relevant questions concern the issuer, claim, validity, current status, intended audience, and the receiving service's acceptance policy. A verified attribute should not be interpreted as unrestricted organizational authority or permission to transact.

Hybrid-ID connects readers to the broader Trust Center and identity API capabilities, including supported privacy proofs. A workflow should request the evidence needed for its specific decision and preserve access controls around the underlying private records. Planned challenge or presentation contracts remain distinct from callable operations. A proof label does not establish that every identity claim can be proved today or that its acceptance conditions are the same across applications.

People, organizations, and the agents they authorize

A person's sign-in can establish a relationship with an application, but the application still needs to determine which organization and workspace that person may represent. The intended Hybrid-ID agent model extends this idea: identify the person or business behind an agent, then establish the agent's own scoped authority for its task. Unified agent access is still being developed. Existing workload bindings and policy capabilities are the integration foundation, not a claim that signing in automatically provisions an agent.

An assistant may be permitted to read an approved context collection and prepare a purchasing proposal while remaining unable to approve or execute it. Private source access, publication to an agent audience, proposal evaluation, and financial execution have separate permissions. The identity relationship makes responsibility easier to explain; it does not collapse those boundaries or copy private business memory between applications.

Illustrative workflow: returning to a purchasing decision

A buyer returns to a participating procurement application through Hybrid-ID. After validating the sign-in, the application restores the buyer's existing workspace and checks their current organizational permissions. A separately authorized workload can retrieve reviewed supplier context for a purchasing assistant. The assistant prepares a proposal, and a permitted policy evaluation can explain whether its exact terms fit the current rules.

StepWhat the workflow establishesSeparate condition
Sign in with Hybrid-IDThe validated identity of the returning participant.The application's workspace and resource permissions.
Apply application permissionsWhich business context and task the participant may access.The workload's own binding and approved capabilities.
Read permitted contextWhich reviewed material can inform the assistant's proposal.Access to private source records and any onward disclosure.
Authorize a specific actionThe required permission and review for the exact operation.Actual signing, execution, and confirmation under the enabled service.

The buyer reviews the proposed terms and supporting references before any separately authorized action. AI Wallet Control's public decision-only preview does not reserve funds, sign, or send the payment. This scenario illustrates the relationship between existing interfaces and application responsibilities; it is not a claim that Hybrid-ID provides a single automatic sign-in-to-payment journey or that unified agent access is already released.

Current availability and architectural direction

AreaPublicly described availabilityHow to evaluate it
Human sign-inOpenID Connect is live for registered applications; Hybrid-Chain uses Hybrid-ID sign-in.Confirm application registration, callbacks, scopes, consent, and the canonical issuer.
New accounts and applicationsPublic account registration is invitation-only; self-service client registration is not available.Agree the approved onboarding route instead of inferring access from an endpoint listing.
Credentials and privacy proofsOptional capabilities have operation-specific scopes and availability; some proof contracts are planned.Check the exact implemented interface and the claim the recipient needs.
Unified agent accessThe integrated identity experience for agents remains in development.Use confirmed workload and policy integrations without assuming automatic delegation.
Decentralized identityVerifiable trust across applications is the direction; current account access uses an operated identity service.Assess the issuer, status, and acceptance of a particular credential rather than assuming a fully decentralized sign-in system.

This availability summary was checked against Hybrid-ID's public product and developer descriptions and live OIDC discovery on October 3, 2026. The linked integration reference remains the place to confirm current contracts and permissions. The architectural advantage is an understandable starting point for connected business: familiar human access can lead into useful context and controlled commerce while each service retains responsibility for the authority it grants.