CSA Research Publications
Whitepapers, Reports and Other Resources
Browse Publications
 | Zero Trust Microsegmentation Guidance Release Date: 09/08/2026 Microsegmentation is a foundational Zero Trust strategy that strengthens security by enforcing explicit, fine-grained communication controls between systems,... Request to download |
 | Cryptographic Context Injection: When Encryption Defeats AI Guardrails Release Date: 08/22/2026 Key Takeaways A new attack technique called cryptographic context injection encrypts malicious instructions with a strong cipher — AES-256-GCM with PBKDF2-de... Request to download |
 | Top Threats to Cloud Computing Survey Report 2026 The 2026 survey results indicate a significant shift in cloud security priorities. Identity, artificial intelligence, third-party dependencies, and APIs now ... Request to download |
| AISI: Open-Weight Models Close Cyber Capability Gap Release Date: 08/09/2026 Key Takeaways The UK AI Security Institute (AISI) has found that the leading open-weight AI models now trail frontier closed models on offensive cyber tasks ... Request to download |
| Four AI Escapes: A Systemic Governance Risk Reading Release Date: 08/09/2026 What Sandbox and Containment Failures at OpenAI and Anthropic Reveal About Evaluation Governance. Key Takeaways Between July 21 and July 30, 2026, OpenAI and... Request to download |
| EU AI Act Article 50: Transparency Obligations Take Effect Release Date: 07/28/2026 Learn what the EU AI Act's Article 50 transparency requirements mean for AI providers and deployers. This publication explains the new disclosure obligations... Request to download |
 | AI Security Through the CISO Lens Release Date: 07/28/2026 Gain insights from CSA's AI Storm Summit series on how CISOs are preparing for AI-driven cyber risks. Explore evolving strategies for AI governance, vulnerab... Request to download |
 | Hugging Face Incident Initial Post-Mortem Release Date: 07/27/2026 Examine CSA’s analysis of the first publicly documented autonomous AI attack. Learn how the incident unfolded, why traditional defenses fell short, and the p... Request to download |
| Every Frontier Model Cheated: What AISI's Findings Mean for Trust Release Date: 07/26/2026 Key Takeaways The UK AI Security Institute (AISI) reported on July 21, 2026 that every one of the five frontier models it evaluated for cybersecurity capabil... Request to download |
 | Release Date: 07/22/2026 A growing population of non-human identities (NHIs) are powering cloud platforms, SaaS applications, automation, and AI-based systems. Traditional Identit... Request to download |
| Hugging Face's Autonomous AI Agent Breach Release Date: 07/19/2026 Security Implications and Guidance for Agentic-Attacker Incidents. Key Takeaways Hugging Face disclosed on July 16, 2026 that an intrusion into its productio... Request to download |
| Agent Data Injection: A New Attack Class Beyond Prompt Injection Release Date: 07/16/2026 How Corrupted Metadata Bypasses Existing Agent Defenses. Agent Data Injection: A New Attack Class Beyond Prompt Injection Key Takeaways Researchers from Seou... Request to download |
| Gold Eagle: The White House's AI Vulnerability Clearinghouse Release Date: 07/15/2026 What the New Federal Coordination Model Means for Critical Infrastructure Operators. Key Takeaways The White House announced Gold Eagle on July 15, 2026, a f... Request to download |
 | Preparing Your Networks for the "AI Storm" Release Date: 07/01/2026 AI is changing the economics of cyber defense. AI models are improving their abilities to discover novel vulnerabilities, create new exploits, and engage ... Request to download |
| Operationalizing the Agentic Control Plane: Integrating the AARM Specification Release Date: 06/22/2026 The CSA Agentic Control Plane (ACP) defines the governance framework through which AI agents are orchestrated, monitored, and controlled across a ten-laye... Request to download |
| Agent Shared Security and Safety Responsibility Model (Agent 3SRM) Release Date: 06/22/2026 As AI agents transition from experimental prototypes to production enterprise systems, the question of who is responsible for what becomes critical. The C... Request to download |
| Operationalizing the Agentic Control Plane: Integrating the Agentic Trust Framework Release Date: 06/22/2026 The CSA Agentic Control Plane (ACP) defines the governance framework through which Agent Owners orchestrate and govern AI agents, and this paper shows how... Request to download |
| Release Date: 06/22/2026 Agentic AI systems mark a fundamental shift from passive intelligence to active agency—autonomously perceiving, reasoning, planning, and executing actions... Request to download |
| Securing the Agentic Control Plane Release Date: 06/22/2026 As AI agents transition from experimental prototypes to production enterprise systems, securing their lifecycle—from creation through runtime operation to... Request to download |
| AI Agents: Architecture and Control Plane Release Date: 06/22/2026 AI agents introduce architectural and security challenges that extend beyond traditional application engineering. Their autonomous operation, tool access,... Request to download |