diff --git a/SCOPE.md b/SCOPE.md
index 9a125fa..9ffecc4 100644
--- a/SCOPE.md
+++ b/SCOPE.md
@@ -21,7 +21,7 @@ NetKingdom is a self-optimizing security platform for Kubernetes-based IT infras
 ## In Scope
 
 - NetKingdom IAM Profile specification (versioned OIDC/PKCE contract;
-  canonical spec: `canon/standards/iam-profile_v0.2.md`)
+  canonical spec: `canon/standards/iam-profile_v0.3.md`)
 - SSO/MFA Platform: Keycloak with LDAP/Entra federation, enterprise identity (NK-WP-0001, finished)
 - Local Identity: file-based user store + minimal OIDC server for bootstrap phase (NK-WP-0002, finished)
 - User Engine Boundary Contract: source-of-truth, membership,
diff --git a/canon/standards/iam-profile_v0.3.md b/canon/standards/iam-profile_v0.3.md
new file mode 100644
index 0000000..546bf49
--- /dev/null
+++ b/canon/standards/iam-profile_v0.3.md
@@ -0,0 +1,473 @@
+---
+id: netkingdom-iam-profile
+type: standard
+title: "NetKingdom IAM Profile v0.3"
+domain: netkingdom
+status: accepted
+version: "0.3"
+created: "2026-07-23"
+updated: "2026-07-23"
+scope: core-platform
+supersedes:
+  - canon/standards/iam-profile_v0.2.md
+adr:
+  - docs/adr/ADR-0011-iam-profile-ownership-and-version-governance.md
+  - docs/adr/ADR-0013-tenant-onboarding-grouping-taxonomy.md
+  - docs/adr/ADR-0014-tenant-capability-roles-and-tenant-engine-ownership.md
+---
+
+# NetKingdom IAM Profile v0.3
+
+> Minor version. Per ADR-0011's versioning rule, this adds an optional claim
+> and clarifies non-normative guidance — no required claim, validation rule,
+> or previously-issued token is invalidated. Existing v0.2 implementations
+> remain conformant; `tenant_roles` and the revised Tenant Claim guidance are
+> additive.
+
+## Purpose
+
+The NetKingdom IAM Profile is the provider-neutral OIDC contract that
+identity implementations issue and applications consume.
+
+It defines:
+
+- OIDC discovery and endpoint requirements;
+- Authorization Code + PKCE for human login;
+- service-account and workload identity token requirements;
+- human, service, and agent principal representation;
+- tenant, tenant-grouping, and platform-boundary claims;
+- tenant capability roles and their carrying mechanism;
+- explicit assurance evidence;
+- the identity-to-authorization claim contract consumed by flex-auth;
+- local-development and emergency-access behavior;
+- executable conformance expectations.
+
+Applications target this profile, not a concrete identity provider.
+key-cape is the lightweight implementation. Keycloak is the expanded-mode
+implementation. Both are interchangeable at the application and
+authorization boundary when they conform to this document.
+
+## Ownership
+
+NetKingdom owns the core/platform profile. See ADR-0011.
+
+Downstream systems may define extension scopes, roles, resource names,
+and tenant policy vocabularies. Those extensions are not part of the core
+profile unless a future version explicitly adopts them. Extension
+vocabularies must map back to the core claims in this document before
+flex-auth or applications consume them.
+
+## Design Principles
+
+- Consumers trust signed OIDC tokens, not provider-specific sessions.
+- Identity providers assert identity and authentication evidence; they do
+  not make final resource authorization decisions.
+- The same profile works in lightweight key-cape mode and expanded
+  Keycloak mode.
+- Tenancy is explicit. `tenant:platform` is distinct from tenant planes
+  such as `tenant:coulomb` and from later tenants grouped per ADR-0013.
+- A tenant's onboarding **grouping** (ADR-0013) and its **capability
+  roles** (ADR-0014) are independent axes. Neither is encoded in the other;
+  a tenant's roles may change without renaming its identifier.
+- Human, service, and agent principals are distinguishable.
+- Assurance evidence is explicit enough for flex-auth policy.
+- Local-development issuers are useful but never accepted by production.
+- Emergency access is auditable, time-bounded, and reviewable.
+
+## Discovery Contract
+
+Every IAM Profile implementation MUST expose OIDC discovery at:
+
+```text
+GET <issuer>/.well-known/openid-configuration
+```
+
+The discovery response MUST include:
+
+| Field | Requirement |
+| --- | --- |
+| `issuer` | Exact issuer identifier used in tokens |
+| `authorization_endpoint` | Required for human Authorization Code + PKCE |
+| `token_endpoint` | Required for token exchange and service accounts |
+| `jwks_uri` | Required for signature validation |
+| `userinfo_endpoint` | Required when userinfo is supported by the flow |
+| `scopes_supported` | MUST include `openid`; SHOULD include `profile` and `email` |
+| `response_types_supported` | MUST include `code` |
+| `grant_types_supported` | MUST include `authorization_code`; MUST include `client_credentials` or a documented workload-token exchange for service identities |
+| `id_token_signing_alg_values_supported` | MUST include the implementation signing algorithm; RS256 is required for v0.2+ conformance |
+| `code_challenge_methods_supported` | MUST include `S256` |
+
+The response SHOULD include `end_session_endpoint` where logout is
+supported and `claims_supported` listing the core claims below.
+
+Consumers MUST discover endpoints and key material from the issuer
+metadata instead of hardcoding provider-specific paths.
+
+## Required Flows
+
+### Human Interactive Flow
+
+Human users authenticate with Authorization Code + PKCE.
+
+Required properties:
+
+- PKCE with `S256` is mandatory for browser and CLI clients.
+- Implicit flow is not part of the profile.
+- MFA or equivalent strong assurance is mandatory for privileged,
+  destructive, platform-root, and emergency access in production.
+- Access tokens are short-lived.
+- Refresh tokens are allowed only for trusted clients with explicit
+  rotation and revocation.
+
+### Service Account Flow
+
+Service-to-service traffic uses client credentials or a deployment's
+documented workload identity token-exchange equivalent.
+
+Required properties:
+
+- Service subjects are stable and named for service plus environment.
+- Secrets or workload credentials are delivered through the
+  credential-management standard, not plaintext configuration.
+- Tokens include an audience that identifies the target service.
+- Tokens carry `principal_type: service`.
+- Service accounts receive only required scopes and roles.
+- Credentials are rotated and never shared between environments.
+
+### Agent Principal Flow
+
+Agents are automation principals that may act autonomously or under
+delegated authority.
+
+Required properties:
+
+- Tokens carry `principal_type: agent`.
+- Tokens include an `agent` object with `id` and `mode`.
+- `agent.mode` is `autonomous` or `delegated`.
+- Delegated agents MUST identify the delegating actor using `actor_sub`
+  or an equivalent `act.sub` claim.
+- Agent tokens MUST carry the tenant they operate within.
+- Agent tokens MUST include assurance evidence for both the agent
+  credential and any delegated human authority when policy needs it.
+
+## Core Claims
+
+Access tokens accepted by production consumers MUST provide the following
+claims after provider mapping or normalization:
+
+| Claim | Type | Meaning |
+| --- | --- | --- |
+| `iss` | string | OIDC issuer URL or issuer identifier |
+| `sub` | string | Stable subject identifier unique within `iss` |
+| `aud` | string or array | Intended audience; MUST include the receiving service |
+| `exp` | number | Expiry timestamp |
+| `iat` | number | Issued-at timestamp |
+| `nbf` | number | Not-before timestamp, recommended for production tokens |
+| `jti` | string | Token identifier, recommended for audit and replay controls |
+| `tenant` | string | Tenant identifier such as `tenant:platform` or `tenant:friendly:binky` |
+| `principal_type` | string | `human`, `service`, or `agent` |
+| `groups` | array | Group memberships, possibly empty |
+| `roles` | array | Coarse **identity** roles for the subject, possibly empty |
+| `scope` or `scp` | string or array | Granted OAuth scopes |
+| `assurance` | object | Authentication and credential assurance evidence |
+
+Recommended human claims:
+
+| Claim | Meaning |
+| --- | --- |
+| `preferred_username` | Human-readable username |
+| `email` | Contact identity |
+| `name` | Display name |
+
+Recommended service claims:
+
+| Claim | Meaning |
+| --- | --- |
+| `azp` or `client_id` | Authorized client/service identifier |
+| `service` | Object naming the service and environment |
+
+Recommended agent claims:
+
+| Claim | Meaning |
+| --- | --- |
+| `agent.id` | Stable agent identifier |
+| `agent.mode` | `autonomous` or `delegated` |
+| `actor_sub` or `act.sub` | Delegating subject for delegated agents |
+
+Optional claims (new in v0.3):
+
+| Claim | Type | Meaning |
+| --- | --- | --- |
+| `tenant_roles` | array | Cached tenant **capability** roles (`PLTF`/`IAM`/`VEN`/`CUS`), possibly empty. See "Tenant Roles" below — this is a point-in-time cache, not the authoritative source. |
+
+### Role Claim
+
+The canonical **subject**-level role claim is `roles`, an array of strings.
+This is distinct from `tenant_roles` (below) — a subject's own coarse
+identity roles are not the same fact as which capability roles the
+subject's tenant holds.
+
+Expanded-mode Keycloak deployments may also expose provider-native roles
+such as `realm_access.roles`, but conforming tokens consumed by flex-auth
+or applications MUST either emit `roles` directly or pass through a
+normalizing adapter that produces `roles`.
+
+### Scope Vocabulary
+
+The core profile defines only OAuth/OIDC base scopes:
+
+| Scope | Meaning |
+| --- | --- |
+| `openid` | Required for OIDC login |
+| `profile` | Basic profile claims |
+| `email` | Email claim where appropriate |
+| `offline_access` | Refresh-token capable access where explicitly allowed |
+
+Hub-, application-, and resource-specific scopes such as `hub:*`,
+`ops:*`, `fin:*`, or storage actions are downstream extensions. They are
+valid only when the consuming system defines them and maps them to
+flex-auth resource/action semantics.
+
+## Tenant Claim
+
+`tenant` is required for every token accepted by profile consumers.
+
+Tenant identifiers follow `tenant:<grouping>:<name>`, where `<grouping>` is
+one of the taxonomy ratified by ADR-0013:
+
+```text
+trial        - test/trial/showcase tenants only
+friendly     - known, easily reached, tolerant of experimentation/instability
+single       - one-person business entities (freelance consultants)
+small        - up to 10 employees at time of onboarding (attoo)
+medium       - up to 100 employees (attoo)
+large        - up to 1000 employees (attoo)
+enterprise   - 1001+ employees (attoo)
+consumer     - private individuals
+family       - a legal family
+community    - a non-legal group of people
+association  - a legal association of people
+agentic      - financially enabled AI entities
+```
+
+`tenant:platform` and `tenant:coulomb` remain **reserved, ungrouped**
+identifiers outside this taxonomy: `tenant:platform` is the platform
+control-plane tenant, not a business entity being onboarded; `tenant:coulomb`
+is the first internal/reference tenant established by ADR-0006, predating
+this taxonomy. Tenant administration for `tenant:coulomb` or any grouped
+tenant must never imply platform-root authority.
+
+Subjects may have access to multiple tenants, but a token used for a
+request MUST identify the tenant context for that request. If a client
+needs to switch tenant context, it obtains a new token or uses an
+approved token-exchange flow that records the target tenant.
+
+The grouping segment is **onboarding-risk / entity-shape** classification
+only. It does not gate which capability roles (below) a tenant may hold —
+see Tenant Roles.
+
+## Tenant Roles
+
+Tenant **capability roles** are a separate fact from the grouping above and
+from the subject-level `roles` claim: `PLTF`, `IAM`, `VEN`, `CUS`
+(ratified by ADR-0014), non-exclusive — a tenant may hold several
+simultaneously, and holding one does not require or restrict any grouping.
+
+**Source of truth:** `tenant-engine`
+(`canon/standards/tenant-engine-boundary-contract_v0.1.md`), not this
+profile and not any token. `tenant-engine` records role grants/revocations,
+their link (if any) to a plan/subscription, and emits domain events on
+change.
+
+**Carrying mechanism — hybrid, not claim-only:**
+
+- `key-cape` (or Keycloak) MAY stamp a cached `tenant_roles` claim onto an
+  issued token at issuance time, sourced from `tenant-engine`.
+- Consumers MAY trust the cached claim for ordinary, non-privileged
+  decisions.
+- Consumers **MUST NOT** trust the cached claim for privileged, destructive,
+  platform-root, secret, credential-vending, or otherwise `assurance.level
+  >= aal2`-class decisions. Those decisions **MUST** query `tenant-engine`
+  live for current role state before authorizing the action.
+- This bounds staleness for ordinary actions to the issuing token's
+  lifetime (5-30 minutes for service/agent tokens, see Token Lifecycle)
+  while guaranteeing freshness exactly where a stale grant (e.g. `VEN`
+  surviving a plan cancellation) would matter most.
+
+`trial`-grouped tenants may hold any capability role without restriction —
+the grouping exists to showcase, test, and explore every role. Safety for
+`trial` tenants is enforced through `tenant-engine`-owned resource
+guardrails (spend limits, entity/action counts — reserved, not yet
+specified), not through role gating.
+
+## Assurance Evidence
+
+The canonical assurance claim is `assurance`.
+
+It is an object with these fields:
+
+| Field | Type | Meaning |
+| --- | --- | --- |
+| `level` | string | `aal0`, `aal1`, `aal2`, `aal3`, or `break_glass` |
+| `methods` | array | Authentication methods, e.g. `pwd`, `otp`, `webauthn`, `client_secret`, `workload_identity`, `upstream_mfa` |
+| `mfa` | boolean | Whether the authentication included multiple factors or equivalent upstream evidence |
+| `source` | string | Provider of the evidence, e.g. `key-cape`, `keycloak`, `privacyidea`, `entra`, `local-identity` |
+| `at` | number | Authentication time, recommended |
+
+Level meanings:
+
+| Level | Meaning |
+| --- | --- |
+| `aal0` | Local/dev or unauthenticated bootstrap evidence; never production privileged |
+| `aal1` | Single-factor or service credential evidence |
+| `aal2` | MFA or equivalent strong upstream assurance |
+| `aal3` | Phishing-resistant or hardware-backed assurance |
+| `break_glass` | Time-bounded emergency access with post-event review |
+
+Privileged, destructive, platform-root, secret, credential-vending, and
+emergency flows require `aal2` or stronger unless a policy explicitly
+permits a narrower service or workload identity path. Emergency access
+MUST use `break_glass` and short token lifetimes. This is the same
+threshold class that gates when `tenant_roles` MUST be re-validated live
+against `tenant-engine` rather than trusted from the token — see Tenant
+Roles.
+
+Provider-native claims such as `acr` and `amr` may be present, but
+consumers use `assurance` as the normalized profile claim.
+
+## Identity To Authorization Contract
+
+flex-auth consumes IAM Profile tokens as normative identity input.
+flex-auth MUST NOT re-derive identity, tenant, group, role, or assurance
+facts from provider-specific session state.
+
+The profile guarantees these inputs for authorization decision envelopes:
+
+| Decision input | Source claim |
+| --- | --- |
+| Subject | `sub` |
+| Issuer | `iss` |
+| Audience | `aud` |
+| Tenant | `tenant` |
+| Principal type | `principal_type` |
+| Groups | `groups` |
+| Subject roles | `roles` |
+| Tenant capability roles | `tenant_roles` (cached); `tenant-engine` live lookup required for high-stakes decisions |
+| Scopes | `scope` or `scp` |
+| Assurance | `assurance` |
+| Authorized client | `azp` or `client_id`, where present |
+| Agent/delegation context | `agent`, `actor_sub`, or `act`, where present |
+| Token lifetime/audit ids | `iat`, `nbf`, `exp`, `jti`, where present |
+
+Authorization decisions are made by flex-auth and its delegated PDP
+adapters. Identity providers may assert roles or scopes, but those claims
+are inputs to policy, not final permission to act on a resource.
+
+## Token Lifecycle
+
+Recommended production defaults:
+
+| Token | Lifetime | Notes |
+| --- | --- | --- |
+| Human access token | 5-15 minutes | Short-lived bearer token |
+| Refresh token | 8-12 hours | Rotated and revoked on logout or suspicion |
+| Service token | 5-30 minutes | Reissued by client credentials or workload identity |
+| Agent token | 5-30 minutes | Shorter when delegated or platform-scoped |
+| Emergency token | 5-15 minutes | Requires incident/review record |
+
+Consumers MUST reject expired tokens and tokens with invalid issuer,
+audience, signature, `nbf`, or algorithm. Clock skew tolerance SHOULD be
+small, normally no more than 60 seconds.
+
+JWKS material may be cached, but consumers MUST tolerate key rotation by
+refreshing JWKS when a token uses an unknown `kid`.
+
+## Local Development Profile
+
+A local file-backed provider may be used for development, tests, and
+bootstrap contexts where the full platform is unavailable.
+
+It MUST:
+
+- expose OIDC discovery;
+- issue signed JWTs;
+- support deterministic test users and service accounts;
+- use local-only issuer URLs or a clearly local issuer identifier;
+- mark tokens as local/development through issuer, audience, or
+  assurance evidence;
+- be rejected by production consumers.
+
+Production consumers MUST reject:
+
+- issuer `local-identity`;
+- `http://` issuers;
+- loopback issuers such as `localhost` or `127.0.0.1`;
+- tokens with `assurance.level: aal0`;
+- tokens where the environment marks the issuer as local/dev.
+
+## Emergency And Break-Glass Access
+
+Emergency access is allowed only as a break-glass path.
+
+Requirements:
+
+- Emergency identities are disabled by default.
+- Activation requires an incident, decision, or human-recorded review
+  reference.
+- Tokens are short-lived and carry the `emergency` role.
+- Tokens carry `assurance.level: break_glass`.
+- Every emergency action emits an audit/progress/incident event.
+- Emergency access is reviewed after use and then disabled again.
+
+Emergency access MUST NOT bypass audit logging or flex-auth policy.
+
+## Conformance
+
+An implementation conforms to IAM Profile v0.3 when it passes the
+executable conformance suite in:
+
+```text
+tools/iam-profile-conformance/
+```
+
+The suite validates:
+
+- discovery document completeness;
+- PKCE `S256` advertisement and rejection of authorization requests that
+  omit a code challenge;
+- JWKS structure and key ids;
+- token issuer, audience, expiry, `nbf`, `iat`, and RS256 signature;
+- tenant, principal type, groups, roles, scopes, and assurance claim
+  shape;
+- `tenant_roles` claim shape when present (array of ratified role strings);
+- agent and delegated-agent claim shape;
+- local-development issuer rejection in production mode.
+
+Conformance must be runnable against both key-cape lightweight issuers
+and Keycloak expanded-mode issuers. Implementations may add provider
+adapters, but the token consumed by applications and flex-auth must match
+the core claim contract above. `tenant_roles` conformance does not require
+an implementation to emit the claim (it is optional); when emitted, it
+must match the ratified vocabulary.
+
+## Validation Checklist
+
+A service or implementation is profile-ready when:
+
+- it reads OIDC discovery rather than hardcoding endpoints;
+- it validates issuer, audience, expiry, `nbf`, algorithm, and
+  signature;
+- it refreshes JWKS on unknown `kid`;
+- it supports Authorization Code + PKCE for human login;
+- it supports service-account or workload identity tokens;
+- it emits `tenant`, `principal_type`, `groups`, `roles`,
+  `scope`/`scp`, and `assurance`;
+- it uses the ADR-0013 grouping vocabulary for new tenant identifiers;
+- if it consumes `tenant_roles`, it treats the claim as a cache and
+  re-validates live against `tenant-engine` before any `aal2`-class
+  decision;
+- it maps provider-native claims into the canonical core claims;
+- it rejects local-development issuers in production;
+- it logs emergency access with a durable audit trail;
+- flex-auth receives identity facts from the profile, not from
+  provider-specific sessions.
diff --git a/canon/standards/tenant-engine-boundary-contract_v0.1.md b/canon/standards/tenant-engine-boundary-contract_v0.1.md
new file mode 100644
index 0000000..51bedaf
--- /dev/null
+++ b/canon/standards/tenant-engine-boundary-contract_v0.1.md
@@ -0,0 +1,163 @@
+---
+id: netkingdom-tenant-engine-boundary-contract
+type: standard
+title: "NetKingdom Tenant Engine Boundary Contract v0.1"
+domain: netkingdom
+status: accepted
+version: "0.1"
+created: "2026-07-23"
+updated: "2026-07-23"
+scope: tenant-domain-boundaries
+adr:
+  - docs/adr/ADR-0006-recursive-multi-tenant-identity-authorization.md
+  - docs/adr/ADR-0013-tenant-onboarding-grouping-taxonomy.md
+  - docs/adr/ADR-0014-tenant-capability-roles-and-tenant-engine-ownership.md
+related:
+  - canon/standards/iam-profile_v0.3.md
+  - canon/standards/user-engine-boundary-contract_v0.1.md
+  - canon/standards/credential-management_v0.2.md
+  - docs/platform-identity-security-architecture.md
+---
+
+# NetKingdom Tenant Engine Boundary Contract v0.1
+
+## Purpose
+
+This contract defines how `tenant-engine` integrates into the NetKingdom
+landscape without duplicating identity, authorization, user/membership,
+pricing, deployment, or secret-custody responsibilities that other services
+already own.
+
+`tenant-engine` owns **tenant-as-an-entity** facts: existence, identifier
+and grouping (ADR-0013), capability roles (ADR-0014), plan/subscription
+assignment, and — reserved, not yet implemented — guardrail/quota policy.
+It is not an identity provider, not a policy decision point, not a billing
+processor, not a pricing-model engine, and not the user/membership domain.
+NetKingdom owns this cross-repo boundary contract; `tenant-engine` and its
+adapters must preserve the ownership rules below even in local/standalone
+deployments.
+
+## Ownership Model
+
+| Concern | Owner | tenant-engine responsibility |
+| --- | --- | --- |
+| OIDC discovery, login, token issuance | `key-cape`, Keycloak | Supply `tenant_roles` via a read API at issuance time; never issues tokens itself |
+| Tenant identifier format and grouping vocabulary | NetKingdom (`iam-profile_v0.3.md`, ADR-0013) | Store and validate tenant records against the ratified vocabulary; does not redefine it |
+| Tenant capability-role vocabulary | NetKingdom (ADR-0014) | Store role grants against the ratified vocabulary (`PLTF`/`IAM`/`VEN`/`CUS`); does not redefine it |
+| Users, profiles, memberships | `user-engine` | References `tenant_id` as an external key only; never stores user or membership records |
+| Fine-grained authorization decisions | `flex-auth` | Serves as an authoritative data source `flex-auth` queries live for high-stakes checks; makes no authorization decisions itself |
+| Pricing-model / plan definitions | `adaptive-pricing` | References plan definitions by id; never forks or duplicates plan terms locally |
+| Runtime secrets | OpenBao / `secrets-engine` | Tenant-scoped OpenBao paths (e.g. `tenants/<name>/...`) reference `tenant-engine`'s tenant registry as the naming authority; `tenant-engine` never stores secret material |
+| Deployment mechanics | Railiance | Publishes readiness signals only |
+| Credential routing/discovery | `ops-warden` | `tenant-engine`'s own service credentials (if any) are routed the same way as any other lane; `tenant-engine` is not a routing authority |
+
+## Source-Of-Truth Matrix
+
+| Resource kind | Source of truth | tenant-engine relation | Boundary rule |
+| --- | --- | --- | --- |
+| Tenant record (existence, grouping, name/slug) | `tenant-engine` | Canonical owner | Grouping value must be one of ADR-0013's taxonomy, or the reserved ungrouped `platform`/`coulomb` identifiers |
+| Tenant capability roles | `tenant-engine` | Canonical owner | Non-exclusive; grant/revoke are audited mutations (see Grant Contract below), never free-form field updates |
+| Plan/subscription assignment | `tenant-engine` | Canonical owner | References an `adaptive-pricing` plan id; does not define plan terms |
+| Pricing-model / plan definitions | `adaptive-pricing` | Consume by reference only | `tenant-engine` must not cache plan terms beyond what's needed to resolve which roles a plan currently grants |
+| Guardrail/quota policy (spend limits, entity/action counts) | `tenant-engine` (reserved) | Canonical owner once designed | Not implemented by this contract version — reserved namespace only, see Guardrail Policy section |
+| `tenant_roles` token claim | `tenant-engine` (live) / `key-cape` (cached copy) | `tenant-engine` is authoritative; the token claim is a point-in-time cache | Privileged/high-stakes decisions MUST re-query `tenant-engine` live — see IAM Profile v0.3, "Tenant Roles" section — never trust the cached claim alone |
+| Tenant identifier claim shape (`tenant`) | NetKingdom / IAM Profile contract | `tenant-engine` validates against it; does not mint the claim format | The wire format stays owned by the profile; `tenant-engine` owns which concrete tenant values currently exist |
+| User/membership records scoped by a tenant | `user-engine` | No relation | `tenant_id` is the only key shared between the two services; `tenant-engine` never stores or reads user data |
+| Authorization decisions | `flex-auth` | Data source only | `tenant-engine` never enforces access itself; it answers queries `flex-auth`'s policy packages issue |
+
+## Tenant Role & Plan Grant Contract
+
+Every role grant or revocation is an audited mutation, not a direct field
+write, and must carry:
+
+| Field | Requirement |
+| --- | --- |
+| `tenant_id` | Stable tenant identifier |
+| `role` | One of `PLTF`, `IAM`, `VEN`, `CUS` |
+| `grant_reason` | `plan_assignment`, `manual_grant`, or `platform_default` |
+| `plan_id` | `adaptive-pricing` plan id; required when `grant_reason: plan_assignment`, otherwise null |
+| `granted_by` | Subject or system that performed the grant |
+| `granted_at` | Timestamp |
+| `revoked_at` | Nullable timestamp |
+| `correlation_id` | Request/import correlation id |
+
+Rules:
+
+- Plan-linked roles (starting with `IAM`) are granted and revoked
+  automatically as a consequence of plan-assignment changes in
+  `tenant-engine`; manual grants are the exception path and must always
+  record `granted_by`.
+- `trial`-grouped tenants may hold any role without a plan requirement —
+  `grant_reason: platform_default` is valid for trial-eligible roles with no
+  `plan_id` (ADR-0014).
+- Every grant or revocation emits a domain event. After a revocation,
+  `key-cape`'s previously-cached `tenant_roles` claim is considered stale
+  until the token naturally expires — v0.1 relies on the profile's short
+  token lifetimes for bounded staleness, not push invalidation.
+- Cross-tenant grants are rejected; a grant always targets exactly one
+  `tenant_id`.
+
+## Authorization Contract
+
+`tenant-engine` is a data source, not a policy enforcement point.
+`flex-auth` remains the policy decision point for authorization outcomes;
+`key-cape` remains the token issuer. `tenant-engine` exposes:
+
+- a **read API** `key-cape` calls at token-issuance time to source the
+  cached `tenant_roles` claim;
+- a **live-lookup API** `flex-auth` calls for high-stakes decisions
+  (`assurance.level >= aal2`-class actions, per IAM Profile v0.3), returning
+  current roles, grouping, and plan state;
+- a **write API**, authorization-gated by `flex-auth` itself, for grant/
+  revoke/plan-assignment mutations — `tenant-engine` does not self-authorize
+  its own writes.
+
+Performance model:
+
+- the live-lookup API must be fast and highly available, since `flex-auth`
+  calls it synchronously on the request path for privileged actions; slow
+  or unavailable `tenant-engine` lookups fail closed (deny), never fail open;
+- the cached-claim read API used by `key-cape` may be slower, since it only
+  runs at token issuance, not per-request;
+- request-scoped memoization within a single `flex-auth` decision is
+  allowed; cross-request caching of live-lookup results is not, since it
+  would recreate the staleness risk the live path exists to avoid.
+
+## Guardrail Policy (Reserved, Not Implemented)
+
+Named here so the shape exists before it's designed, matching ADR-0014:
+
+- Spend limits — `trial`-grouped tenants default to a **zero** spend budget.
+- Entity count limits — reserved, not yet specified.
+- Action count limits — reserved, not yet specified.
+- Enforcement point, override process, and how guardrails interact with
+  capability roles (a `trial` tenant holding `VEN` is still bound by `trial`
+  guardrails) are all open design questions for `tenant-engine`'s own
+  workplan, not resolved by this contract.
+
+## Audit Correlation Contract
+
+Every `tenant-engine` mutation (tenant creation, role grant/revoke, plan
+assignment) must create or reference a correlation bundle:
+
+| Field | Requirement |
+| --- | --- |
+| `correlation_id` | Stable id propagated across request, decision, audit, and event records |
+| `actor` | IAM Profile actor envelope (who performed the mutation) |
+| `tenant_id` | Tenant affected |
+| `resource` and `action` | e.g. `tenant-engine:role`, `grant` / `revoke` |
+| `authorization_decision_id` | `flex-auth` decision id backing the write |
+| `outbox_event_id` | Event id emitted for the mutation |
+| `change_summary` | Redacted mutation summary |
+
+## Non-Goals
+
+- `tenant-engine` does not process payments or hold financial instruments —
+  it records which plan a tenant is assigned to, sourced from whatever
+  system actually processes payment (not yet identified; out of scope for
+  this contract).
+- `tenant-engine` does not define pricing models, tiers, or terms —
+  `adaptive-pricing` does.
+- `tenant-engine` does not store user accounts, profiles, or memberships —
+  `user-engine` does.
+- `tenant-engine` does not make authorization decisions — `flex-auth` does.
diff --git a/docs/adr/ADR-0014-tenant-capability-roles-and-tenant-engine-ownership.md b/docs/adr/ADR-0014-tenant-capability-roles-and-tenant-engine-ownership.md
new file mode 100644
index 0000000..81d03c9
--- /dev/null
+++ b/docs/adr/ADR-0014-tenant-capability-roles-and-tenant-engine-ownership.md
@@ -0,0 +1,171 @@
+# ADR-0014 - Tenant Capability Roles, Carrying Mechanism, and Tenant-Engine Ownership
+
+**Status:** Accepted
+**Date:** 2026-07-23
+**Deciders:** Bernd Worsch, Codex
+
+## Context
+
+ADR-0013 introduced the tenant onboarding grouping taxonomy
+(`trial`/`friendly`/`single`/.../`agentic`), deliberately orthogonal to a
+separate, unratified **capability-role** model sketched in
+`docs/princedom-isolation-exploration.md`: `PLTF` (operates the platform),
+`IAM` (organizes its own users/auth/secrets), `VEN` (provides apps/services
+to others), `CUS` (consumes apps/services from `PLTF` or `VEN` tenants) —
+non-exclusive, a tenant may hold several at once.
+
+That exploration left open where capability roles actually live (a per-token
+claim vs. a registry), who owns them, how they're granted or revoked, and
+how this interacts with the IAM Profile's existing `roles` claim — which is
+a *per-subject* claim ("coarse identity roles" for the human/service/agent
+holding the token), a different concept from a *per-tenant* capability fact.
+Conflating the two would be a category error: `roles: ["VEN"]` on a token
+would ambiguously mean "this subject has vendor-role" vs. "this subject's
+tenant is a vendor."
+
+No existing service owns tenant-as-an-entity facts (existence, grouping,
+capability roles, plan/subscription state) as a queryable resource.
+`user-engine`'s own boundary contract
+(`canon/standards/user-engine-boundary-contract_v0.1.md`) explicitly scopes
+`user-engine` to *consuming* tenant identifiers and *storing tenant-scoped
+records*, not owning tenant identity or capability facts.
+
+Bernd's direction (2026-07-23):
+
+- Implement the previously-discussed hybrid carrying approach: cache a
+  tenant's capability roles on the token for ordinary decisions, but
+  require a live check for critical/high-stakes actions.
+- Role grants are usually tied to a payment plan — most concretely, `IAM`
+  means the tenant has its own dedicated key-cape/Keycloak instance for
+  isolation, scale, and performance, which is itself a paid capability.
+- `trial`-grouped tenants may hold **any** capability role without
+  restriction, specifically so the platform can showcase, test, and explore
+  every role. Safety for trial tenants comes from resource guardrails
+  (spend limits defaulting to zero budget, entity/action count limits), not
+  from role gating — guardrail design itself is future work, not this ADR.
+- A new service, `tenant-engine`, will be built (Bernd) as the owner of this
+  domain, organized beside `user-engine` rather than inside it — smaller,
+  single-purpose services are easier to reason about and drift less, matching
+  the fleet's existing convention (`activity-core`, `audit-core`, `user-engine`,
+  and others).
+
+## Decision
+
+1. **Capability-role vocabulary ratified as core NetKingdom vocabulary**:
+   `PLTF`, `IAM`, `VEN`, `CUS`, non-exclusive. `IAM` specifically means: the
+   tenant operates its own dedicated IAM implementation instance (lightweight
+   key-cape or expanded Keycloak) rather than sharing the platform's, for
+   isolation/scale/performance — not "any tenant that happens to have users."
+
+2. **`tenant-engine` is the canonical owner** of tenant-domain facts: tenant
+   existence, grouping (ADR-0013), capability roles (this ADR), plan/
+   subscription assignment, and — reserved for future design, not built now —
+   guardrail/quota policy. It is a new, separate service, not a module inside
+   `user-engine`. Its ownership boundary is defined in the companion contract,
+   `canon/standards/tenant-engine-boundary-contract_v0.1.md`.
+
+3. **Carrying mechanism: hybrid cache + live re-validation.** `tenant-engine`
+   is the single source of truth. `key-cape` stamps a cached, optional
+   `tenant_roles` claim onto issued tokens at issuance time, sourced from
+   `tenant-engine` (added to the IAM Profile as a new optional claim —
+   `canon/standards/iam-profile_v0.3.md`). Consumers may trust the cached
+   claim for ordinary decisions. `flex-auth` **MUST** re-validate live
+   against `tenant-engine` — never trust the cached claim alone — before
+   authorizing privileged or high-stakes actions, using the same threshold
+   class the profile already defines for `assurance.level >= aal2`
+   (privileged, destructive, platform-root, secret, credential-vending
+   flows). This bounds staleness risk for ordinary actions to a token's
+   short lifetime (5–30 minutes for service/agent tokens, per the profile's
+   Token Lifecycle table) while guaranteeing freshness exactly where it
+   matters most.
+
+4. **Role governance is plan-linked.** Granting a role is normally a
+   consequence of a tenant's payment-plan state in `tenant-engine`, not a
+   separate manual workflow — starting with `IAM`. `tenant-engine` records
+   which plan grants which role(s); `adaptive-pricing` remains the source of
+   plan/pricing-model *definitions*, `tenant-engine` owns the tenant's
+   *current* plan/subscription assignment, referenced by id, never
+   duplicated locally. Whether `VEN` needs an approval gate beyond payment
+   (reselling access carries legal/compliance exposure a payment alone
+   doesn't cover) is **not resolved by this ADR** — left to `tenant-engine`'s
+   own workplan.
+
+5. **Trial tenants may hold any capability role, unrestricted.** The `trial`
+   grouping's purpose (showcase, test, explore) requires demonstrating every
+   role. Safety is enforced through resource guardrails instead: trial
+   tenants default to a **spend budget of zero**, with entity and action
+   count limits to follow. Guardrail policy design (exact limits,
+   enforcement point, override process) is real, near-term future work,
+   reserved as `tenant-engine`'s to own once designed — not specified by
+   this ADR.
+
+6. **Grouping and capability role are independent axes** recorded on the
+   same tenant record in `tenant-engine`. Neither constrains the other
+   except where a future guardrail policy explicitly says so.
+
+## Consequences
+
+- `canon/standards/iam-profile_v0.3.md` adds the optional `tenant_roles`
+  claim and folds in ADR-0013's tenant-identifier vocabulary update (both
+  non-breaking per ADR-0011's own minor-version rule — no existing
+  implementation is invalidated by either change). Supersedes
+  `iam-profile_v0.2.md`.
+- `tenant-engine` becomes a new repository with its own workplans (Bernd).
+  `canon/standards/tenant-engine-boundary-contract_v0.1.md` defines its
+  ownership boundary now, before code exists — the same sequencing
+  `user-engine`'s contract followed.
+- `flex-auth` policy packages gating high-stakes actions must add a
+  `tenant-engine` live-lookup step; they cannot trust `tenant_roles` alone
+  for those decisions.
+- `key-cape` needs a `tenant-engine` integration at token-issuance time to
+  source the cached claim — tracked in `key-cape`'s own workplans, not here.
+- `docs/platform-identity-security-architecture.md`'s Tenant Model section
+  is updated to reflect the grouping + role split and `tenant-engine`'s role
+  (companion change alongside this ADR).
+- Guardrail/quota policy is named as required near-term work and given an
+  owner (`tenant-engine`), but is explicitly not designed by this ADR.
+
+## Alternatives Considered
+
+### Token-claim-only, no live re-validation
+
+Rejected: staleness would be unbounded within a token's lifetime for
+genuinely high-stakes actions. A stale `VEN` grant surviving a plan
+downgrade or cancellation is not an acceptable risk for money-movement or
+credential-vending flows — exactly the class the profile already treats as
+requiring the strongest assurance.
+
+### Registry-only, no cached claim
+
+Rejected: every ordinary request would pay a `tenant-engine` round-trip even
+for non-critical checks, adding latency and a hard runtime dependency for
+every consumer, not just the ones handling privileged actions.
+
+### Restrict role eligibility by tenant grouping (e.g., `trial` cannot hold `VEN`)
+
+Rejected per Bernd's direction: `trial` tenants exist specifically to
+showcase every role. Resource guardrails are the intended safety mechanism
+instead, keeping the two axes (grouping, role) independent.
+
+### Put tenant-role/plan storage inside `user-engine`
+
+Rejected. `user-engine`'s own boundary contract scopes it to consuming
+tenant identifiers and storing tenant-*scoped* user records, not owning
+tenant-as-an-entity facts. A dedicated service avoids coupling a
+security-critical, high-frequency lookup (used by `flex-auth` on every
+privileged decision, and by `key-cape` on every token issuance) to
+`user-engine`'s much larger surface (registration flows, factor models,
+family dataspace onboarding) that has nothing to do with tenant capability
+state.
+
+## Follow-Up
+
+- `tenant-engine` repository creation and its own workplan (Bernd).
+- `key-cape` integration: source `tenant_roles` from `tenant-engine` at
+  token issuance.
+- `flex-auth` policy package updates: live `tenant-engine` re-validation
+  gate for privileged actions.
+- Guardrail/quota policy design for `trial` (and eventually all) tenants:
+  spend limits, entity/action count limits, enforcement point, override
+  process.
+- Resolve whether `VEN` needs an approval gate beyond payment-plan state.
diff --git a/docs/platform-identity-security-architecture.md b/docs/platform-identity-security-architecture.md
index 161c4e7..03f18d4 100644
--- a/docs/platform-identity-security-architecture.md
+++ b/docs/platform-identity-security-architecture.md
@@ -1,7 +1,7 @@
 # Platform Identity and Security Architecture
 
 Status: implemented architecture baseline for NetKingdom/Railiance/Coulomb
-Date: 2026-05-24
+Date: 2026-05-24 (Tenant Model updated 2026-07-23, ADR-0013/ADR-0014)
 
 ## Purpose
 
@@ -296,15 +296,39 @@ Examples of platform-root actions:
 Every protected resource should belong to a tenant or to the platform
 control plane.
 
-Suggested identifiers:
+Tenant identifiers are `tenant:<grouping>:<name>`, where `<grouping>` is an
+onboarding-risk / entity-shape classification ratified by ADR-0013
+(`trial`, `friendly`, `single`, `small`, `medium`, `large`, `enterprise`,
+`consumer`, `family`, `community`, `association`, `agentic`):
 
 ```text
-tenant:platform          # platform control plane resources
-tenant:coulomb           # first internal/reference tenant
-tenant:sandbox:<name>    # sandbox tenants
-tenant:customer:<name>   # future customer tenants
+tenant:platform            # platform control plane — reserved, ungrouped
+tenant:coulomb              # first internal/reference tenant — reserved, ungrouped
+tenant:friendly:binky       # first tenant onboarded under this taxonomy
+tenant:<grouping>:<name>    # general shape for tenants onboarded going forward
 ```
 
+`tenant:platform` and `tenant:coulomb` predate the taxonomy and stay
+reserved outside it — see `canon/standards/iam-profile_v0.3.md`'s Tenant
+Claim section for the full rationale.
+
+A tenant separately holds **capability roles** — `PLTF`, `IAM`, `VEN`, `CUS`
+(ADR-0014), non-exclusive — describing what it *does* on the platform,
+independent of its grouping (what kind of entity it *is*). A `trial`-grouped
+tenant may hold any capability role, since `trial` exists to showcase every
+role; safety instead comes from tenant-level resource guardrails (spend
+limits, entity/action counts — reserved, not yet designed).
+
+`tenant-engine` (new service, `canon/standards/tenant-engine-boundary-contract_v0.1.md`)
+is the canonical owner of both tenant grouping and capability-role facts,
+plus plan/subscription assignment. `key-cape` caches a tenant's roles on
+issued tokens (`tenant_roles` claim) for ordinary decisions; `flex-auth`
+re-validates live against `tenant-engine` before authorizing privileged,
+destructive, platform-root, or otherwise `aal2`-class actions — the same
+threshold the assurance model below already uses. `user-engine` continues
+to own users and memberships *scoped by* a tenant, not the tenant record
+itself.
+
 Tenant membership and platform membership are distinct. A subject may be
 an administrator in `tenant:coulomb` without being a platform operator.