CIS Introduces AI Security Guides Covering Models, Agents and MCP
Summary
CIS and security partners introduced three companion guides extending the CIS Controls to AI systems, with the LLM guide addressing context integrity, sensitive data, deployment choices and model supply-chain security.
Key points
- CIS, Astrix Security and Cequence Security developed three companion guides for AI LLMs, AI agents and the Model Context Protocol (MCP).
- The LLM guide treats model inputs as untrusted and recommends sanitizing retrieved context, hardening system prompts and addressing indirect prompt injection.
- It advises classifying model-related data and using retention, deletion, encryption and access controls to reduce leakage risks.
- The guide distinguishes security obligations for endpoint-hosted, enterprise-hosted and SaaS-hosted models.
- It recommends treating models as software artifacts through provenance tracking, version pinning, integrity checks and retirement policies.
Article Details
- Event Type
- Publication of three CIS Companion Guides for securing AI systems, including the AI Large Language Model (LLM) Companion Guide.
- Impact
- The guides extend the CIS Controls to AI environments and offer guidance on model-input integrity, sensitive-data handling, deployment differences, and model supply-chain provenance. The article describes them as helping organizations reduce AI-related security risks.
Products
AI Agent Companion GuideThe result is three new CIS Companion Guides: the AI Large Language Model (LLM) Companion Guide, the AI Agent Companion Guide, and the Model Context Protocol (MCP) Companion Guide.AI Large Language Model (LLM) Companion GuideThe result is three new CIS Companion Guides: the AI Large Language Model (LLM) Companion Guide, the AI Agent Companion Guide, and the Model Context Protocol (MCP) Companion Guide.CIS ControlsThe work extends the globally recognized CIS Critical Security Controls® (CIS Controls®) into environments where autonomous decision‑making, tool and API access, and automated threats introduce new risks.Model Context Protocol (MCP) Companion GuideThe result is three new CIS Companion Guides: the AI Large Language Model (LLM) Companion Guide, the AI Agent Companion Guide, and the Model Context Protocol (MCP) Companion Guide.