grc.store

Search / finos-ccc/ccc.marefarc.cn / v2026.06-rc1

Release · v2026.06-rc1

FINOS-CCC/CCC.MARefArc.CN Control Catalog

FINOS-CCC/CCC.MARefArc.CN

Controls for Multi-Agent Reference Architecture technologies, as defined by the FINOS Common Cloud Controls project.

Published by FINOS Common Cloud Controls

License No license declared

Install

OCI v1.1
$grcli unpack --repository finos-ccc/ccc.marefarc.cn --version v2026.06-rc1

grcli unpack verifies this signature against the recorded identity below and fails closed before writing any files — no separate verify step needed. Pass --no-verify to skip.

Coordinate
oci.grc.store/finos-ccc/ccc.marefarc.cn:v2026.06-rc1
Manifest digest
sha256:eaa5b9bb87a0cefb172f6b9305ce155dd8ddb6060be97eeff2f330b8e02f52f0
Signed by
no signature recorded

Identity recorded by this hub when the version was published; unpack (above) checks the signature against it.

Verify in CI — check the signature without downloading
$grcli verify --repository finos-ccc/ccc.marefarc.cn --version v2026.06-rc1

Read-only: downloads nothing and exits non-zero if the signature or the recorded identity doesn't match — use it as an admission/policy gate.

Provenance

1 layer
Digest Media type Size
5f3f04b5da5f… application/vnd.gemara.artifact.v1+yaml 28.7 KiB
Bundle config blob 
{
  "bundle-version": "1.0",
  "gemara-version": "v1.2.0",
  "metadata": {
    "provenance": {
      "buildDefinition": {
        "buildType": "https://grc.store/grcli/buildtype/v0",
        "externalParameters": {
          "artifact": {
            "id": "CCC.MARefArc.CN",
            "type": "ControlCatalog"
          },
          "target": {
            "registry": "oci.grc.store",
            "repository": "finos-ccc/ccc.marefarc.cn",
            "tag": "v2026.06-rc1"
          }
        },
        "internalParameters": {
          "CI": "true",
          "GITHUB_ACTIONS": "true",
          "GITHUB_ACTOR": "eddie-knight",
          "GITHUB_REF": "refs/heads/main",
          "GITHUB_REPOSITORY": "eddie-knight/common-cloud-controls",
          "GITHUB_RUN_ATTEMPT": "1",
          "GITHUB_RUN_ID": "27958973490",
          "GITHUB_SHA": "65b6844e7d4361034aac75cde94963f5e9d73820",
          "GITHUB_WORKFLOW": "Release Catalog",
          "RUNNER_OS": "Linux"
        },
        "resolvedDependencies": [
          {
            "name": "artifacts/ai-ml/multi-agent-refarch/controls.yaml",
            "uri": "file://artifacts/ai-ml/multi-agent-refarch/controls.yaml",
            "digest": {
              "sha256": "5f3f04b5da5f2903b1646893828be2789dd3ae2aa0dbce98c92ba5396407b570"
            }
          },
          {
            "name": "source",
            "uri": "git+https://github.com/eddie-knight/common-cloud-controls@65b6844e7d4361034aac75cde94963f5e9d73820",
            "digest": {
              "gitCommit": "65b6844e7d4361034aac75cde94963f5e9d73820"
            }
          }
        ]
      },
      "runDetails": {
        "builder": {
          "id": "https://github.com/eddie-knight/common-cloud-controls/actions/runs/27958973490",
          "version": {
            "go": "go1.25.0",
            "go-arch": "amd64",
            "go-os": "linux",
            "grcli": "v0.2.2"
          }
        },
        "metadata": {
          "invocationId": "27958973490-1",
          "startedOn": "2026-06-22T14:11:24.644675405Z",
          "finishedOn": "2026-06-22T14:11:24.838693963Z"
        },
        "byproducts": [
          {
            "name": "controls.yaml",
            "digest": {
              "sha256": "5f3f04b5da5f2903b1646893828be2789dd3ae2aa0dbce98c92ba5396407b570"
            }
          }
        ]
      }
    }
  },
  "artifacts": [
    {
      "name": "controls.yaml",
      "type": "ControlCatalog",
      "id": "CCC.MARefArc.CN",
      "role": "artifact"
    }
  ]
}

CCC Multi-Agent Reference Architecture Controls

Controls for Multi-Agent Reference Architecture technologies, as defined by the FINOS Common Cloud Controls project.

ID
CCC.MARefArc.CN
Version
v2026.06-rc1
Gemara version
v1.2.0
Author
FINOS Common Cloud Controls

Preventive

Prevention and Risk Mitigation

  1. CCC.MARefArc.CN01 Data Filtering From External Knowledge Bases

    Objective

    Sanitize, filter, and classify data ingested by the Knowledge Layer from internal and external source bases before it is embedded into the vector store or used for retrieval-augmented generation, preventing inadvertent exposure or manipulation of sensitive organizational knowledge.

    Assessment requirements

    1. Data ingested into the Knowledge Layer MUST be scanned and filtered for sensitive content before it is embedded or indexed for retrieval.

      Applicability: all

    2. Ingestion pipelines MUST enforce source-level allow and deny rules so that unapproved repositories cannot be embedded into the vector store.

      Applicability: all

    References to Other Documents
    Guidelines
    • finos-air
      • AIR-PREV-002
    Threats
    • CCC.MARefArc.Threats
      • CCC.MARefArc.TH06
      • CCC.MARefArc.TH07
      • CCC.MARefArc.TH01
      • CCC.MARefArc.TH02

  2. CCC.MARefArc.CN02 User, Application, and Model Firewalling

    Objective

    Establish enforced trust boundaries between the user, the application, and the models and tools by routing all traffic through the agent, LLM, and MCP gateways where guardrails inspect and constrain requests and responses.

    Assessment requirements

    1. All traffic between the user-facing application and the agent layer MUST traverse the Agent Gateway guardrails; direct application-to-agent paths that bypass the gateway MUST be denied.

      Applicability: all

    2. All model and tool invocations MUST traverse the LLM and MCP gateway guardrails respectively, with inputs and outputs inspected against policy.

      Applicability: all

    References to Other Documents
    Guidelines
    • finos-air
      • AIR-PREV-003
    Threats
    • CCC.MARefArc.Threats
      • CCC.MARefArc.TH08
      • CCC.MARefArc.TH09
      • CCC.MARefArc.TH10
      • CCC.MARefArc.TH14
      • CCC.MARefArc.TH15
      • CCC.MARefArc.TH11
      • CCC.MARefArc.TH12
      • CCC.MARefArc.TH13

  3. CCC.MARefArc.CN03 System Acceptance Testing

    Objective

    Validate agents, models, and end-to-end workflows against accuracy, robustness, bias, drift, and compliance criteria before promotion to production, and re-validate after material changes.

    Assessment requirements

    1. Each agent and model configuration MUST pass a documented acceptance test suite covering accuracy, bias and fairness, and compliance criteria before being onboarded into the respective registry.

      Applicability: all

    2. Acceptance testing MUST be repeated when a pinned model version, system prompt, or deployment configuration changes.

      Applicability: all

    References to Other Documents
    Guidelines
    • finos-air
      • AIR-PREV-005
    Threats
    • CCC.MARefArc.Threats
      • CCC.MARefArc.TH23
      • CCC.MARefArc.TH25
      • CCC.MARefArc.TH19
      • CCC.MARefArc.TH16
      • CCC.MARefArc.TH17
      • CCC.MARefArc.TH18

  4. CCC.MARefArc.CN04 Data Quality and Classification

    Objective

    Assess the quality of, and assign classification and sensitivity labels to, all data used for grounding, training, and fine-tuning, and enforce handling rules derived from those labels throughout the Knowledge and LLM layers.

    Assessment requirements

    1. All data sources feeding the Knowledge Layer or used for fine-tuning MUST carry a classification and sensitivity label.

      Applicability: all

    2. Data quality checks for accuracy, completeness, freshness, and bias MUST be applied to grounding and training data, and data failing thresholds MUST be quarantined from production use.

      Applicability: all

    References to Other Documents
    Guidelines
    • finos-air
      • AIR-PREV-006
    Threats
    • CCC.MARefArc.Threats
      • CCC.MARefArc.TH06
      • CCC.MARefArc.TH07
      • CCC.MARefArc.TH22
      • CCC.MARefArc.TH23
      • CCC.MARefArc.TH25
      • CCC.MARefArc.TH26
      • CCC.MARefArc.TH01
      • CCC.MARefArc.TH02
      • CCC.MARefArc.TH03
      • CCC.MARefArc.TH04
      • CCC.MARefArc.TH05
      • CCC.MARefArc.TH16

  5. CCC.MARefArc.CN05 Legal and Contractual Frameworks for AI Systems

    Objective

    Establish contractual controls with model and MCP service providers covering data handling, retention and deletion, intellectual property, liability, and supply-chain integrity.

    Assessment requirements

    1. Every approved model and MCP server provider MUST be governed by a contract specifying data handling, retention and deletion, and intellectual-property terms before it is added to the registry.

      Applicability: all

    2. Provider contracts MUST include supply-chain integrity and breach-notification obligations.

      Applicability: all

    References to Other Documents
    Guidelines
    • finos-air
      • AIR-PREV-007
    Threats
    • CCC.MARefArc.Threats
      • CCC.MARefArc.TH25
      • CCC.MARefArc.TH26
      • CCC.MARefArc.TH01
      • CCC.MARefArc.TH02
      • CCC.MARefArc.TH20
      • CCC.MARefArc.TH21
      • CCC.MARefArc.TH15

  6. CCC.MARefArc.CN06 Quality of Service and DDoS Prevention

    Objective

    Protect model and tool availability by enforcing quality-of-service controls, rate limits, and abuse and DDoS mitigation at the gateways.

    Assessment requirements

    1. The LLM and MCP gateways MUST enforce per-consumer rate limits and quotas.

      Applicability: all

    2. Gateways MUST apply DDoS and abuse detection and load shedding to preserve availability under load.

      Applicability: all

    References to Other Documents
    Guidelines
    • finos-air
      • AIR-PREV-008
    Threats
    • CCC.MARefArc.Threats
      • CCC.MARefArc.TH08
      • CCC.MARefArc.TH09
      • CCC.MARefArc.TH10

  7. CCC.MARefArc.CN07 AI Model Version Pinning

    Objective

    Pin and record explicit model versions in the Model Registry so that model behaviour is reproducible and provider-side changes are surfaced rather than silently absorbed.

    Assessment requirements

    1. The Model Registry MUST pin an explicit, immutable model version for each approved model used in production.

      Applicability: all

    2. A change to the underlying provider model version MUST require re-approval through the registry before use.

      Applicability: all

    References to Other Documents
    Guidelines
    • finos-air
      • AIR-PREV-010
    Threats
    • CCC.MARefArc.Threats
      • CCC.MARefArc.TH19
      • CCC.MARefArc.TH17

  8. CCC.MARefArc.CN08 Role-Based Access Control for AI Data

    Objective

    Enforce least-privilege, role-based access control over all AI data stores, including source bases, the vector store, and model artifacts.

    Assessment requirements

    1. Access to AI data stores, including source bases, the vector store, and model artifacts, MUST be governed by role-based access control with least privilege.

      Applicability: all

    2. Access grants MUST be reviewed periodically and revoked when no longer required.

      Applicability: all

    References to Other Documents
    Guidelines
    • finos-air
      • AIR-PREV-012
    Threats
    • CCC.MARefArc.Threats
      • CCC.MARefArc.TH06
      • CCC.MARefArc.TH07
      • CCC.MARefArc.TH01
      • CCC.MARefArc.TH02
      • CCC.MARefArc.TH20
      • CCC.MARefArc.TH21

  9. CCC.MARefArc.CN09 Encryption of AI Data at Rest

    Objective

    Encrypt AI data at rest, including the vector store and source repositories, so that storage-level access does not expose embeddings or sensitive content.

    Assessment requirements

    1. The vector store and all AI data repositories MUST be encrypted at rest using managed keys.

      Applicability: all

    2. Access to encryption keys MUST be restricted and audited.

      Applicability: all

    References to Other Documents
    Guidelines
    • finos-air
      • AIR-PREV-014
    Threats
    • CCC.MARefArc.Threats
      • CCC.MARefArc.TH25
      • CCC.MARefArc.TH03
      • CCC.MARefArc.TH04
      • CCC.MARefArc.TH05

  10. CCC.MARefArc.CN10 AI Firewall Implementation and Management

    Objective

    Implement and operate an AI firewall within the guardrail components that inspects prompts, content, and responses for injection, sensitive data, and policy violations.

    Assessment requirements

    1. The gateway guardrails MUST include an AI firewall that screens inputs for prompt injection and policy violations and screens outputs for sensitive-data disclosure and harmful content.

      Applicability: all

    2. AI firewall rules MUST be centrally managed and versioned.

      Applicability: all

    References to Other Documents
    Guidelines
    • finos-air
      • AIR-PREV-017
    Threats
    • CCC.MARefArc.Threats
      • CCC.MARefArc.TH08
      • CCC.MARefArc.TH09
      • CCC.MARefArc.TH10
      • CCC.MARefArc.TH14
      • CCC.MARefArc.TH15
      • CCC.MARefArc.TH11
      • CCC.MARefArc.TH12
      • CCC.MARefArc.TH13

  11. CCC.MARefArc.CN11 Agent Authority Least Privilege Framework

    Objective

    Constrain each agent's authority to the minimum set of tools, APIs, and data required for its task, enforced by the runtime and MCP guardrails, and prevent permission creep during operation.

    Assessment requirements

    1. Each agent MUST be granted only the tools, endpoints, and scopes required for its registered purpose.

      Applicability: all

    2. Agent permissions MUST be enforced at invocation time and MUST NOT expand during a session without explicit reauthorization.

      Applicability: all

    References to Other Documents
    Guidelines
    • finos-air
      • AIR-PREV-018
    Threats
    • CCC.MARefArc.Threats
      • CCC.MARefArc.TH27
      • CCC.MARefArc.TH14

  12. CCC.MARefArc.CN12 Tool Chain Validation and Sanitization

    Objective

    Validate tool selection, sanitize tool-call parameters, and constrain tool sequencing within the runtime and MCP guardrails to prevent manipulation of agent tool use.

    Assessment requirements

    1. Tool-call parameters MUST be validated and sanitized against a schema before execution.

      Applicability: all

    2. Tool selection and sequencing MUST be constrained to patterns authorized for the agent's task.

      Applicability: all

    References to Other Documents
    Guidelines
    • finos-air
      • AIR-PREV-019
    Threats
    • CCC.MARefArc.Threats
      • CCC.MARefArc.TH27
      • CCC.MARefArc.TH28
      • CCC.MARefArc.TH11
      • CCC.MARefArc.TH12
      • CCC.MARefArc.TH13

  13. CCC.MARefArc.CN13 MCP Server Security Governance

    Objective

    Govern the onboarding, verification, and ongoing monitoring of MCP servers so that only approved, integrity-verified servers are reachable, and supply-chain compromise is detected.

    Assessment requirements

    1. Only MCP servers registered and verified in the MCP Server Registry MUST be reachable by agents.

      Applicability: all

    2. MCP server updates MUST be integrity-verified, and connections MUST be authenticated and transport-encrypted.

      Applicability: all

    References to Other Documents
    Guidelines
    • finos-air
      • AIR-PREV-020
    Threats
    • CCC.MARefArc.Threats
      • CCC.MARefArc.TH29
      • CCC.MARefArc.TH01
      • CCC.MARefArc.TH02
      • CCC.MARefArc.TH20
      • CCC.MARefArc.TH21

  14. CCC.MARefArc.CN14 Multi-Agent Isolation and Segmentation

    Objective

    Isolate agents and their memory and state so that compromise or failure of one agent cannot propagate to others, and enforce segmentation of agent-to-agent communication.

    Assessment requirements

    1. Each agent's runtime, memory, and workspace MUST be isolated such that one agent cannot read or modify another's state without authorization.

      Applicability: all

    2. Agent-to-agent communication MUST be authenticated and segmented to prevent privilege inheritance and cascade failures.

      Applicability: all

    References to Other Documents
    Guidelines
    • finos-air
      • AIR-PREV-022
    Threats
    • CCC.MARefArc.Threats
      • CCC.MARefArc.TH27
      • CCC.MARefArc.TH30
      • CCC.MARefArc.TH31

  15. CCC.MARefArc.CN15 Agentic System Credential Protection Framework

    Objective

    Prevent agents from discovering, extracting, or misusing credentials by brokering secrets outside agent-accessible surfaces and constraining tool access to credential stores.

    Assessment requirements

    1. Credentials MUST be brokered by the runtime and MUST NOT be exposed to agent-accessible memory, prompts, logs, or the workspace file system.

      Applicability: all

    2. Agent tools MUST be denied access to credential stores, environment secrets, and instance metadata except through the secure-execution broker.

      Applicability: all

    References to Other Documents
    Guidelines
    • finos-air
      • AIR-PREV-023
    Threats
    • CCC.MARefArc.Threats
      • CCC.MARefArc.TH27
      • CCC.MARefArc.TH29
      • CCC.MARefArc.TH32

Detective

Detection and Continuous Improvement

  1. CCC.MARefArc.CN16 AI Data Leakage Prevention and Detection

    Objective

    Detect leakage of sensitive data in model inputs and outputs and in telemetry, and alert and respond when disclosure is detected.

    Assessment requirements

    1. Model outputs and telemetry MUST be monitored for disclosure of classified or sensitive data, with alerts raised on detection.

      Applicability: all

    2. Detected leakage events MUST trigger a documented response and escalation workflow.

      Applicability: all

    References to Other Documents
    Guidelines
    • finos-air
      • AIR-DET-001
    Threats
    • CCC.MARefArc.Threats
      • CCC.MARefArc.TH01
      • CCC.MARefArc.TH02

  2. CCC.MARefArc.CN17 AI System Observability

    Objective

    Instrument every layer to emit logs, traces, metrics, and events to the Observability Layer so that behaviour, drift, availability, and data handling are continuously visible and auditable.

    Assessment requirements

    1. Every layer MUST emit logs, traces, metrics, and events covering requests, model interactions, retrievals, and policy decisions to the Observability Layer.

      Applicability: all

    2. Telemetry MUST be retained and queryable for the period required to support audit and incident investigation.

      Applicability: all

    References to Other Documents
    Guidelines
    • finos-air
      • AIR-DET-004
    Threats
    • CCC.MARefArc.Threats
      • CCC.MARefArc.TH22
      • CCC.MARefArc.TH01
      • CCC.MARefArc.TH02
      • CCC.MARefArc.TH08
      • CCC.MARefArc.TH09
      • CCC.MARefArc.TH10
      • CCC.MARefArc.TH19
      • CCC.MARefArc.TH17
      • CCC.MARefArc.TH18
      • CCC.MARefArc.TH14

  3. CCC.MARefArc.CN18 AI System Alerting and Denial of Wallet Monitoring

    Objective

    Monitor spend and usage of models and tools, and alert on anomalous consumption indicative of Denial of Wallet or runaway agentic loops.

    Assessment requirements

    1. Model and tool consumption MUST be metered per consumer and monitored against budget and rate thresholds.

      Applicability: all

    2. Anomalous spend or call volume MUST raise alerts and MUST be able to trigger throttling or suspension.

      Applicability: all

    References to Other Documents
    Guidelines
    • finos-air
      • AIR-DET-009
    Threats
    • CCC.MARefArc.Threats
      • CCC.MARefArc.TH08
      • CCC.MARefArc.TH09
      • CCC.MARefArc.TH10

  4. CCC.MARefArc.CN19 Human Feedback Loop for AI Systems

    Objective

    Capture human feedback on agent outputs through the Feedback Engine and Human Supervision capabilities and feed it into evaluation and improvement of agents and models.

    Assessment requirements

    1. The system MUST provide a mechanism for humans to rate, correct, or reject agent outputs, and MUST capture that feedback in the Feedback Engine.

      Applicability: all

    2. Captured feedback MUST be reviewable and MUST inform retraining, prompt, or configuration changes.

      Applicability: all

    References to Other Documents
    Guidelines
    • finos-air
      • AIR-DET-011
    Threats
    • CCC.MARefArc.Threats
      • CCC.MARefArc.TH23
      • CCC.MARefArc.TH19
      • CCC.MARefArc.TH17
      • CCC.MARefArc.TH18
      • CCC.MARefArc.TH15

  5. CCC.MARefArc.CN20 Citations and Source Traceability for AI-Generated Information

    Objective

    Attach citations and source traceability to AI-generated information so that outputs can be verified against retrieved sources and decisions can be explained.

    Assessment requirements

    1. Outputs grounded in retrieved content MUST include citations identifying the source documents.

      Applicability: all

    2. The system MUST retain the linkage between an output and the retrieved sources used to produce it.

      Applicability: all

    References to Other Documents
    Guidelines
    • finos-air
      • AIR-DET-013
    Threats
    • CCC.MARefArc.Threats
      • CCC.MARefArc.TH24
      • CCC.MARefArc.TH25
      • CCC.MARefArc.TH16
      • CCC.MARefArc.TH15

  6. CCC.MARefArc.CN21 Automated Evaluation Using LLM-as-a-Judge

    Objective

    Use automated model-based evaluation in the Evaluation Layer to assess output quality, grounding, bias, and policy compliance at scale.

    Assessment requirements

    1. Agent outputs MUST be subject to automated model-based evaluation for grounding, accuracy, and policy compliance.

      Applicability: all

    2. Automated evaluation results MUST feed runtime protection and human supervision when thresholds are breached.

      Applicability: all

    References to Other Documents
    Guidelines
    • finos-air
      • AIR-DET-015
    Threats
    • CCC.MARefArc.Threats
      • CCC.MARefArc.TH22
      • CCC.MARefArc.TH23
      • CCC.MARefArc.TH01
      • CCC.MARefArc.TH02
      • CCC.MARefArc.TH19
      • CCC.MARefArc.TH16
      • CCC.MARefArc.TH17
      • CCC.MARefArc.TH18

  7. CCC.MARefArc.CN22 Preserving Source Data Access Controls in AI Systems

    Objective

    Propagate the access controls of source data into the retrieval path so that retrieval and generation cannot expose content a requesting user is not authorized to see.

    Assessment requirements

    1. Retrieval MUST enforce the access controls of the underlying source data for the requesting principal.

      Applicability: all

    2. Embeddings and retrieved context MUST be filtered by the requester's entitlements before being supplied to the model.

      Applicability: all

    References to Other Documents
    Guidelines
    • finos-air
      • AIR-DET-016
    Threats
    • CCC.MARefArc.Threats
      • CCC.MARefArc.TH25
      • CCC.MARefArc.TH03
      • CCC.MARefArc.TH04
      • CCC.MARefArc.TH05

  8. CCC.MARefArc.CN23 Agent Decision Audit and Explainability

    Objective

    Record an auditable trace of agent decisions, including tool selections, inputs, and rationale, sufficient to explain and review autonomous actions after the fact.

    Assessment requirements

    1. The system MUST record an auditable decision trace for each agent action, including the tools invoked and the inputs and rationale leading to them.

      Applicability: all

    2. Decision traces MUST be retained and reviewable to support explainability and post-incident analysis.

      Applicability: all

    References to Other Documents
    Guidelines
    • finos-air
      • AIR-DET-021
    Threats
    • CCC.MARefArc.Threats
      • CCC.MARefArc.TH27
      • CCC.MARefArc.TH28
      • CCC.MARefArc.TH25