Skip to main content

AI/ML / Multi Agent Refarch

Multi-Agent Reference Architecture Controls

Version: DEV

IDTitleObjectiveControl FamilyThreat MappingsGuideline MappingsAssessment Requirements
CCC.MARefArc.CN01Data Filtering From External Knowledge BasesSanitize, 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.PREV
4
1
2
CCC.MARefArc.CN02User, Application, and Model FirewallingEstablish 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.PREV
8
1
2
CCC.MARefArc.CN03System Acceptance TestingValidate 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.PREV
6
1
2
CCC.MARefArc.CN04Data Quality and ClassificationAssess 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.PREV
12
1
2
CCC.MARefArc.CN05Legal and Contractual Frameworks for AI SystemsEstablish contractual controls with model and MCP service providers covering data handling, retention and deletion, intellectual property, liability, and supply-chain integrity.PREV
7
1
2
CCC.MARefArc.CN06Quality of Service and DDoS PreventionProtect model and tool availability by enforcing quality-of-service controls, rate limits, and abuse and DDoS mitigation at the gateways.PREV
3
1
2
CCC.MARefArc.CN07AI Model Version PinningPin 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.PREV
2
1
2
CCC.MARefArc.CN08Role-Based Access Control for AI DataEnforce least-privilege, role-based access control over all AI data stores, including source bases, the vector store, and model artifacts.PREV
6
1
2
CCC.MARefArc.CN09Encryption of AI Data at RestEncrypt AI data at rest, including the vector store and source repositories, so that storage-level access does not expose embeddings or sensitive content.PREV
4
1
2
CCC.MARefArc.CN10AI Firewall Implementation and ManagementImplement and operate an AI firewall within the guardrail components that inspects prompts, content, and responses for injection, sensitive data, and policy violations.PREV
8
1
2
CCC.MARefArc.CN11Agent Authority Least Privilege FrameworkConstrain 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.PREV
2
1
2
CCC.MARefArc.CN12Tool Chain Validation and SanitizationValidate tool selection, sanitize tool-call parameters, and constrain tool sequencing within the runtime and MCP guardrails to prevent manipulation of agent tool use.PREV
5
1
2
CCC.MARefArc.CN13MCP Server Security GovernanceGovern the onboarding, verification, and ongoing monitoring of MCP servers so that only approved, integrity-verified servers are reachable, and supply-chain compromise is detected.PREV
5
1
2
CCC.MARefArc.CN14Multi-Agent Isolation and SegmentationIsolate 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.PREV
3
1
2
CCC.MARefArc.CN15Agentic System Credential Protection FrameworkPrevent agents from discovering, extracting, or misusing credentials by brokering secrets outside agent-accessible surfaces and constraining tool access to credential stores.PREV
3
1
2
CCC.MARefArc.CN16AI Data Leakage Prevention and DetectionDetect leakage of sensitive data in model inputs and outputs and in telemetry, and alert and respond when disclosure is detected.DET
2
1
2
CCC.MARefArc.CN17AI System ObservabilityInstrument 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.DET
10
1
2
CCC.MARefArc.CN18AI System Alerting and Denial of Wallet MonitoringMonitor spend and usage of models and tools, and alert on anomalous consumption indicative of Denial of Wallet or runaway agentic loops.DET
3
1
2
CCC.MARefArc.CN19Human Feedback Loop for AI SystemsCapture human feedback on agent outputs through the Feedback Engine and Human Supervision capabilities and feed it into evaluation and improvement of agents and models.DET
5
1
2
CCC.MARefArc.CN20Citations and Source Traceability for AI-Generated InformationAttach citations and source traceability to AI-generated information so that outputs can be verified against retrieved sources and decisions can be explained.DET
4
1
2
CCC.MARefArc.CN21Automated Evaluation Using LLM-as-a-JudgeUse automated model-based evaluation in the Evaluation Layer to assess output quality, grounding, bias, and policy compliance at scale.DET
8
1
2
CCC.MARefArc.CN22Preserving Source Data Access Controls in AI SystemsPropagate 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.DET
4
1
2
CCC.MARefArc.CN23Agent Decision Audit and ExplainabilityRecord an auditable trace of agent decisions, including tool selections, inputs, and rationale, sufficient to explain and review autonomous actions after the fact.DET
3
1
2