> ## 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.

# S4: Governance

> AdCP specialist module S4: Governance protocol mastery. Content standards, property lists, campaign governance lifecycle, policy registry, and compliance automation with sandbox agents.

# S4: Governance

> [!NOTE]
>
> **Members only** — Requires Practitioner credential. \~60 minutes with Addie. Combines hands-on lab and adaptive exam.

This specialist module tests your mastery of the full governance protocol. You'll work with sandbox agents to manage content standards, property lists, and collection lists, execute the campaign governance lifecycle (plans, validation, outcomes, audit), resolve compliance policies from the registry, and reason about how governance domains compose. Addie evaluates both your hands-on work and your conceptual understanding.

Passing earns the **AdCP specialist — Governance** credential.

## Specialisms this track prepares you to validate

The following `specialisms` fall under the `governance` domain. Each has its own compliance storyboard — see the [Compliance Catalog](/dist/docs/3.3.0-beta.0/building/verification/compliance-catalog) for the full taxonomy.

| Specialism | Status | What it covers |
| - | - | - |
| `content-standards` | stable | Content standards enforcement (brand safety, policy compliance) |
| `property-lists` | stable | Curated inclusion and exclusion lists for targeting and delivery compliance |
| `collection-lists` | stable | Curated inclusion and exclusion lists of content programs (shows, series, podcasts) |
| `audience-sync` | stable | Syncs buyer-provided audience segments into a platform for activation |
| `governance-delivery-monitor` | stable | Campaign delivery monitoring with drift detection |
| `governance-spend-authority` | stable | Conditional spend approval and human-in-the-loop governance |

## What you'll demonstrate

* Create, update, list, and delete content standards
* Create, update, list, and delete property lists
* Create, update, list, and delete collection lists — program-level brand safety across platforms
* Calibrate content against standards and interpret results
* Create and validate campaign plans with budget authority and policy configuration
* Sync governance agents to accounts via `sync_governance`
* Execute the full governance loop: `sync_plans`, `check_governance` (intent + execution), `report_plan_outcome`
* Use the policy registry to resolve and apply compliance policies
* Interpret audit logs and reason about drift metrics
* Distinguish the three layers of brand safety:
  * **Property lists** — *where* ads run. Stateful lists of properties (domains, apps, CTV services) with include/exclude semantics. Filter supply at the inventory level.
  * **Collection lists** — *what content* ads run in. Shows, series, movies, and sports programs identified by distribution IDs (Gracenote SH/MV/SP). Program-level exclusions independent of which property carries them.
  * **Content standards** — *per-impression adjacency*. Buyer-defined rules that evaluate actual content (not metadata) at serve time via calibrate → execute → validate. Use for episode-level or ephemeral content.
* Explain when to use each layer alone and how they compose (most buyers use one or two)
* Explain how governance domains compose — campaign, property, collection, content standards, creative
* Handle `GOVERNANCE_DENIED` end-to-end: read the `governance_context`, identify the failed rule, correct the payload, and retry
* Walk through the 15-step seller verification of a signed `governance_context` (JWS): `alg` allowlist, `typ` match, JWKS resolution via SSRF-validated fetch, `aud`/opaque-action-`sub`/`phase` binding, `jti` replay dedup, revocation-list check. Explain why each step closes a specific attack (spoofed issuer, cross-action token swap, token replay, key compromise) without disclosing `plan_id` to the seller.
* Apply governance across `purchase_type` values: `media_buy`, `rights_license`, `signal_activation`, `creative_services`. All share the same loop; media-buy-specific validations (channel compliance, seller concentration, pacing) apply only when the payload contains the relevant fields
* Explain the `governance_context` correlation model: governance agent issues an opaque token on first check; buyer attaches it to the media buy envelope; seller echoes it on execution checks so the agent reconnects each lifecycle event without re-deriving state
* Configure audience constraints on a campaign plan using policy categories and restricted attributes
* Validate that `check_governance` correctly denies targeting that violates audience constraints

## Prerequisite reading

### Core governance tasks

> **[property\_lists](/dist/docs/3.3.0-beta.0/governance/property/tasks/property_lists)**
>
> Create and manage property lists for inclusion/exclusion filtering.

> **[collection\_lists](/dist/docs/3.3.0-beta.0/governance/collection/tasks/collection_lists)**
>
> Create and manage collection lists for program-level brand safety.

> **[create\_content\_standards](/dist/docs/3.3.0-beta.0/governance/content-standards/tasks/create_content_standards)**
>
> Define brand-specific content standards for automated compliance.

> **[calibrate\_content](/dist/docs/3.3.0-beta.0/governance/content-standards/tasks/calibrate_content)**
>
> Test whether content meets defined standards before placement.

> **[validate\_property\_delivery](/dist/docs/3.3.0-beta.0/governance/property/tasks/validate_property_delivery)**
>
> Verify that ads were delivered to authorized properties.

> **[sync\_governance](/dist/docs/3.3.0-beta.0/accounts/tasks/sync_governance)**
>
> Sync governance agent endpoints to accounts for seller-side validation.

> **[sync\_plans](/dist/docs/3.3.0-beta.0/governance/campaign/tasks/sync_plans)**
>
> Create and update campaign plans that define authorized parameters.

> **[check\_governance](/dist/docs/3.3.0-beta.0/governance/campaign/tasks/check_governance)**
>
> Validate intended and executed actions against the campaign plan.

> **[report\_plan\_outcome](/dist/docs/3.3.0-beta.0/governance/campaign/tasks/report_plan_outcome)**
>
> Report confirmed outcomes and commit budget against the plan.

> **[get\_plan\_audit\_logs](/dist/docs/3.3.0-beta.0/governance/campaign/tasks/get_plan_audit_logs)**
>
> Review governance decisions, findings, and drift metrics.

### Supporting concepts

> **[Governance overview](/dist/docs/3.3.0-beta.0/governance/overview)**
>
> The governance protocol: property, content standards, creative, and campaign governance.

> **[Campaign governance](/dist/docs/3.3.0-beta.0/governance/campaign)**
>
> Multi-party validation, budget authority, and the trust model.

> **[Campaign governance safety model](/dist/docs/3.3.0-beta.0/governance/campaign/safety-model)**
>
> Three-party trust model: orchestrator, governance agent, and seller.

> **[Campaign governance specification](/dist/docs/3.3.0-beta.0/governance/campaign/specification)**
>
> Full technical specification: validation categories, budget tracking, and audience governance.

> **[Policy registry](/dist/docs/3.3.0-beta.0/governance/policy-registry)**
>
> Community-maintained compliance policies: regulations (COPPA, GDPR, HFSS) and standards (alcohol, pharma).

> **[Property governance](/dist/docs/3.3.0-beta.0/governance/property/index)**
>
> Identity, authorization, and data enrichment for publishers.

> **[Collection governance](/dist/docs/3.3.0-beta.0/governance/collection/index)**
>
> Program-level brand safety with collection lists across CTV, podcast, and other media.

> **[Content standards overview](/dist/docs/3.3.0-beta.0/governance/content-standards/index)**
>
> How content standards work: calibration, local execution, and validation.

> **[Content standards implementation](/dist/docs/3.3.0-beta.0/governance/content-standards/implementation-guide)**
>
> Implementation guide for content standards.

> **[Content artifacts](/dist/docs/3.3.0-beta.0/governance/content-standards/artifacts)**
>
> Artifacts generated during content evaluation.

> **[Creative governance](/dist/docs/3.3.0-beta.0/governance/creative/index)**
>
> Creative quality, compliance, and feature analysis.

> **[Provenance verification](/dist/docs/3.3.0-beta.0/governance/creative/provenance-verification)**
>
> Verifying AI creative provenance and disclosure.

> **[get\_creative\_features](/dist/docs/3.3.0-beta.0/governance/creative/get_creative_features)**
>
> Analyzing creative features for compliance evaluation.

> **[validate\_content\_delivery](/dist/docs/3.3.0-beta.0/governance/content-standards/tasks/validate_content_delivery)**
>
> Post-delivery content validation.

> **[get\_media\_buy\_artifacts](/dist/docs/3.3.0-beta.0/governance/content-standards/tasks/get_media_buy_artifacts)**
>
> Retrieving evaluation artifacts for audit and review.

## Connecting to the test agent

Lab exercises run against the public test agent. Use the shared token — no signup required:

```bash theme={null}
export ADCP_AUTH_TOKEN="1v8tAhASaUYYp4odoQ1PnMpdqNaMiTrCRqYo9OJp6IQ"
export SALES_AGENT_URL="https://test-agent.adcontextprotocol.org/sales/mcp"
export GOVERNANCE_AGENT_URL="https://test-agent.adcontextprotocol.org/governance/mcp"
export ROOT_AGENT_URL="https://test-agent.adcontextprotocol.org/mcp"
```

Use the endpoint that owns the task. The training agent deliberately splits the workflow across tenants:

| Task | Endpoint |
| - | - |
| `sync_governance` | `$SALES_AGENT_URL` |
| `sync_plans`, `check_governance`, `report_plan_outcome`, `get_plan_audit_logs` | `$GOVERNANCE_AGENT_URL` |
| `validate_property_delivery` | `$ROOT_AGENT_URL` (legacy root-only custom tool) |

Do not send `sync_governance` to the governance tenant: it configures the seller account, so the sales tenant owns it. The root route for `validate_property_delivery` is an explicit sandbox compatibility surface, not a claim that every governance task belongs on `/mcp`.

See the [Quickstart](/dist/docs/3.3.0-beta.0/quickstart) for a walkthrough of your first call.

## Lab exercises

1. **Content standards lifecycle** — Create, update, and manage content standards for a fictional brand
2. **Property list management** — Create inclusion and exclusion lists for supply path control
3. **Collection list management** — Create a do-not-air collection list with distribution identifiers, content rating filters, and genre exclusions. Understand how collection lists compose with property lists and content standards for three-layer brand safety.
4. **Content calibration** — Calibrate sample content against your standards and interpret results
5. **Compliance verification** — Verify that sandbox campaign deliveries meet governance requirements
6. **Campaign governance lifecycle** — Sync governance agents via `sync_governance`, create a campaign plan, validate actions (intent + execution), handle denials and conditions, report outcomes, and review audit logs
7. **GOVERNANCE\_DENIED recovery** — Submit a `check_governance` request that fails (audience uses a restricted attribute, or media buy exceeds committed budget). Read the denial reason, correct the payload, and verify the retry passes. Trace `governance_context` through the orchestrator → seller execution check to prove the correlation model.
8. **Purchase type variation** — Run `check_governance` with `purchase_type: "rights_license"` and `purchase_type: "signal_activation"` on the same plan. Observe which validations apply uniformly (budget, geo, flight) and which are media-buy-specific.
9. **Policy resolution and compliance** — Resolve policies from the registry, configure jurisdiction-scoped enforcement, verify that violations are caught
10. **Audience governance** — Configure a plan with `policy_categories` (e.g., `fair_housing`) that carry `restricted_attributes`. Submit `check_governance` requests with audience selectors that use restricted signals and verify the governance agent denies them. Then submit compliant targeting and confirm approval.
11. **Audience drift detection** — Run a series of delivery-phase governance checks with `audience_distribution` indices that gradually shift from baseline parity. Observe how cumulative indices reveal sustained bias patterns that single-period noise obscures, and how governance findings escalate when drift exceeds thresholds.

## Assessment

| Dimension | Weight | What Addie evaluates |
| - | - | - |
| Protocol mastery | 25% | Full governance protocol mastery across all domains including campaign governance lifecycle |
| Safety expertise | 25% | Understands the three-party trust model, separation of duties, and how governance domains compose |
| Oracle understanding | 20% | Understands AI-driven evaluation models and policy registry integration |
| Compliance skill | 30% | Handles regulatory requirements, policy resolution, jurisdiction scoping, and audit interpretation |

Passing threshold: 70%.

## Start this module

> **[Start S4 with Addie](https://agenticadvertising.org/chat)**
>
> "I'd like to start the governance specialist module."


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.