> ## Documentation Index > Fetch the complete documentation index at: https://docs.adcontextprotocol.org/llms.txt > Use this file to discover all available pages before exploring further. # Refresh agent snapshot > Re-probe the agent and update its registry health (online, tools_count, response_time_ms), capability snapshot (inferred type, discovered tools), and compliance verdict (storyboard pass/fail counts). Use after fixing your agent so the registry shows fresh data without waiting for the periodic heartbeat (~1h). **Compliance re-run:** when the caller owns the agent or is an AAO admin and the capability probe succeeds, the full storyboard suite can run for several minutes on capability-rich agents with a fresh test session, and `agent_storyboard_status` is updated. Owner-triggered runs use `triggered_by: 'owner_test'`; admin-triggered support runs use `triggered_by: 'manual'`. Badge fan-out reissues verification badges off the new run. If the compliance call fails (timeout, OAuth wall, internal error), the capability/health portion still returns successfully — `compliance.ran` is `false` with an `error` string. **Auth:** owner of the agent, AAO admin, or static `ADMIN_API_KEY`. **Rate limits:** 60 seconds per agent URL, 30 requests per user per hour. ## OpenAPI ````yaml /static/openapi/registry.yaml post /api/registry/agents/{encodedUrl}/refresh openapi: 3.1.0 info: title: AgenticAdvertising.org Registry API description: >- REST API for the AgenticAdvertising.org registry. Resolve brands, discover properties, look up agents, and validate authorization in the AdCP ecosystem. Most endpoints are public and require no authentication. Endpoints marked with a lock icon accept either an organization API key or a user JWT obtained via the OAuth 2.1 flow — see [Authentication](https://agenticadvertising.org/docs/registry/index#authentication). **Base URL:** `https://agenticadvertising.org` version: 1.0.0 contact: name: AgenticAdvertising.org url: https://agenticadvertising.org servers: - url: https://agenticadvertising.org description: Production security: [] tags: - name: Onboarding description: >- Explicitly bootstrap a third-party integration into the AAO registry. Most callers don't need this tag — `POST /api/me/agents` auto-creates the org (for fresh users) and the member profile (for first-time agent registration) without a separate round trip. Use `POST /api/organizations` only when you need to override the auto-derived org name / company_type / revenue_tier. Tier transitions happen via the billing flow only; the Stripe webhook is the sole writer of `organizations.membership_tier`. - name: Member Agents description: >- Register, list, update, and remove agents on the caller's organization member profile. Authenticated programmatic surface for CI / scripts that don't want to round-trip the full member profile. - name: Brand Resolution description: Resolve advertiser domains to canonical brand identities. - name: Property Resolution description: >- Resolve publisher domains to their property configurations and authorized agents. - name: Agent Discovery description: >- Browse the federated agent network, search agent inventory profiles, publisher index, and registry statistics. - name: Change Feed description: Poll cursor-based registry change events for local sync. - name: Lookups & Authorization description: >- Look up agents by domain or property, and validate ad-serving authorization. - name: Validation Tools description: >- Validate publisher adagents.json files and generate compliant configurations. - name: Community Mirrors description: >- Publish, fetch, list, and retire catalog-only adagents.json mirrors for platforms that have not adopted AdCP. - name: Search description: Cross-entity search across brands, publishers, agents, and properties. - name: Agent Probing description: >- Connect to live agents and inspect their capabilities, formats, and inventory. - name: Brand Discovery description: Discover and crawl brand.json files across domains. - name: Agent Compliance description: Agent compliance status, storyboard test results, and compliance history. - name: Policy Registry description: >- Browse, resolve, and contribute governance policies for campaign compliance. paths: /api/registry/agents/{encodedUrl}/refresh: post: tags: - Agent Compliance summary: Refresh agent snapshot description: >- Re-probe the agent and update its registry health (online, tools_count, response_time_ms), capability snapshot (inferred type, discovered tools), and compliance verdict (storyboard pass/fail counts). Use after fixing your agent so the registry shows fresh data without waiting for the periodic heartbeat (~1h). **Compliance re-run:** when the caller owns the agent or is an AAO admin and the capability probe succeeds, the full storyboard suite can run for several minutes on capability-rich agents with a fresh test session, and `agent_storyboard_status` is updated. Owner-triggered runs use `triggered_by: 'owner_test'`; admin-triggered support runs use `triggered_by: 'manual'`. Badge fan-out reissues verification badges off the new run. If the compliance call fails (timeout, OAuth wall, internal error), the capability/health portion still returns successfully — `compliance.ran` is `false` with an `error` string. **Auth:** owner of the agent, AAO admin, or static `ADMIN_API_KEY`. **Rate limits:** 60 seconds per agent URL, 30 requests per user per hour. operationId: refreshAgent parameters: - schema: type: string description: URL-encoded agent URL example: https%3A%2F%2Fvastlint.org%2Fmcp required: true description: URL-encoded agent URL name: encodedUrl in: path responses: '200': description: Snapshot refreshed content: application/json: schema: type: object properties: online: type: boolean tools_count: type: - integer - 'null' response_time_ms: type: - integer - 'null' inferred_type: type: string description: >- Type inferred from discovered tools (sales, creative, signals, governance, etc.) or 'unknown' type_promoted: type: boolean description: >- True when registry type was upgraded from unknown to inferred_type oauth_required: type: boolean checked_at: type: string error: type: string compliance: type: object properties: ran: type: boolean description: >- True if the full storyboard suite ran and agent_storyboard_status was updated. False when ownership couldn't be resolved, the agent reported auth_required, or the compliance call itself failed. run_id: type: string description: >- Compliance run id written by this refresh. Use with /compliance/diagnostics?run_id=... to inspect failing-step wire evidence. test_session_id: type: string description: >- Fresh test session id used for the compliance run. Useful when matching seller-side logs to the refresh. requested_compliance_target: type: string description: >- Requested compliance target before alias resolution, e.g. 3.1, 3.0, 3.1-rc, or 3.1-beta. Present when `ran` is true. adcp_version: type: string description: >- Concrete AdCP compliance bundle version used for the run, e.g. 3.0.12 or 3.1.0-beta.7. Present when `ran` is true. badge_eligible: type: boolean description: True when this run can update public badge state. badge_eligible_adcp_versions: type: array items: type: string description: >- Public badge versions this run can issue, e.g. ['3.0']. overall_status: type: string description: >- Aggregate verdict from the run (passing / failing / partial / unknown). Only present when `ran` is true. storyboards_passing: type: integer description: Number of storyboards passing on this run. storyboards_total: type: integer description: Number of storyboards evaluated on this run. observations_count: type: integer description: Number of advisory observations emitted by this run. notices_count: type: integer description: Number of run-summary notices emitted by this run. error: type: string description: Reason compliance didn't run when `ran` is false. required: - ran description: >- Compliance re-run summary. The capability/health portion of the response is independent of this block — a failed compliance run still returns the rest of the snapshot. required: - online - tools_count - response_time_ms - inferred_type - type_promoted - oauth_required - checked_at - compliance '400': description: Invalid agent URL content: application/json: schema: $ref: '#/components/schemas/Error' '401': description: Authentication required content: application/json: schema: $ref: '#/components/schemas/Error' '403': description: Not authorized — must be owner or AAO admin content: application/json: schema: $ref: '#/components/schemas/Error' '409': description: Monitoring paused for this agent content: application/json: schema: $ref: '#/components/schemas/Error' '429': description: Rate limit exceeded content: application/json: schema: type: object properties: error: type: string retry_after: type: integer description: Seconds to wait before retrying required: - error - retry_after '502': description: Probe failed (timeout, DNS, OAuth wall, etc.) content: application/json: schema: $ref: '#/components/schemas/Error' security: - bearerAuth: [] - oauth2: [] components: schemas: Error: type: object properties: error: type: string required: - error securitySchemes: bearerAuth: type: http scheme: bearer description: >- Bearer token in the `Authorization` header. Two token types are accepted: - **Organization API key** (`sk_...`) issued via the dashboard. Org-scoped, long-lived, for server-to-server use. - **User JWT** obtained via the OAuth 2.1 authorization code flow with PKCE. User-scoped, short-lived. Discover the authorization server at `/.well-known/oauth-authorization-server` and the protected-resource metadata at `/.well-known/oauth-protected-resource/api`. oauth2: type: oauth2 description: >- OAuth 2.1 authorization code flow with PKCE. Users authenticate via AuthKit and clients receive a Bearer JWT that authorizes both the MCP endpoint and this REST API. Dynamic client registration is supported at `/register`. flows: authorizationCode: authorizationUrl: https://agenticadvertising.org/authorize tokenUrl: https://agenticadvertising.org/token refreshUrl: https://agenticadvertising.org/token scopes: openid: User identifier profile: User profile information email: User email address ````