Security Guidance
CoSAI offers resources and guidelines to implement robust security controls and mitigations, enhancing the trust and security of AI systems within your organization. New papers and frameworks are published regularly.
Download the Papers
- Model Context Protocol (MCP) Security V2.0
- Addendum: Post-Quantum Cryptography Considerations for MCP
- Zero Trust for AI Systems
- AI Shared Responsibility Framework
- Agentic Identity and Access Management
- The Future of Agentic Security: From Chatbots to Autonomous Swarms
- Model Context Protocol (MCP) Security
- AI Incident Response Framework
- Signing ML Artifacts: Building towards tamper-proof ML metadata records
- Preparing Defenders of AI Systems
- Establish Risks and Controls for the AI Supply Chain
You can also access our comprehensive whitepapers and technical documentation on GitHub to dive deeper into each workstream.
