Advertisement
AI Red Teaming: Guardrail Manipulation via Jailbreaking and Data Poisoning
Explores AI red teaming methods like jailbreaking and data poisoning used to manipulate AI guardrails and harden machine learning models against adversarial attacks.
Securing Agentic AI: CISA and International Partners Issue Guidance
CISA and international partners release guidance on securing agentic AI services, detailing risks like autonomous execution and supply chain vulnerabilities.
LiteLLM Proxy Data Exposure & Modification — Urgent Patch Required
Critical vulnerability in LiteLLM proxy enables unauthorized database read/modify access. Exploitation observed shortly after disclosure. Patch immediately.
Secure AI Agent Delegation: Bridging the Authority Gap
AI agents introduce a structural authority gap in enterprise security. Learn how continuous observability serves as a decision engine for delegation.
LMDeploy SSRF: CVE-2026-33626 Exploit and Mitigation Guide
Attackers are actively exploiting CVE-2026-33626, a high-severity SSRF in LMDeploy, to access sensitive LLM data. Learn how to detect and patch this flaw.
Security Risks of Agentic AI in Enterprise Ecosystems
Analysis of security risks in Agentic AI adoption, focusing on prompt injection, autonomous execution, and enterprise mitigation strategies.
Advertisement
CVE-2026-5760: SGLang RCE via Malicious GGUF Models - Patch Now
Critical CVE-2026-5760 command injection in SGLang allows remote code execution via GGUF files. High-performance LLM serving environments are at risk.
Strategic Human-LLM Interaction: Research into AI Trust and Rationality
New research shows humans attribute higher rationality and cooperation to LLMs in strategic games, impacting trust in automated cybersecurity environments.
Google DeepMind Research: Six Web Attack Vectors Against AI Agents
DeepMind researchers reveal how malicious web content can manipulate AI agents, highlighting risks like indirect prompt injection and data exfiltration.
OpenAI Model Behavior Bug Bounty: Reporting AI Safety Risks
OpenAI launches a bug bounty program targeting model abuse and safety risks. Learn how to report jailbreaks and bypasses to improve enterprise AI security.
Governing Agentic AI: Security Risks and Governance Lessons from OpenClaw
Explore the security implications of agentic AI systems like OpenClaw. Learn about the shift to autonomous AI actions and the need for robust governance.
RSAC 2024: AI Security Startups Lead Innovation Sandbox Finalists
Analyze how AI-driven cybersecurity startup trends dominated the 2024 RSAC Innovation Sandbox, signaling a shift toward securing large language models.
Architectural Security Risks of MCP in LLM Environments
Explore architectural security risks introduced by MCP in Large Language Model environments, deemed unpatchable and requiring fundamental redesigns for future safety.
Claudy Day: Prompt Injection and XSS Flaws Target Claude AI Users
Researchers uncover 'Claudy Day', a trio of vulnerabilities in Anthropic's Claude AI that allow data theft through malicious Google search results.
Pentagon CTO and Anthropic Clash Over AI Autonomous Warfare Limits
Pentagon CTO Emil Michael reveals friction with Anthropic over AI safety restrictions hindering the development of autonomous military decision systems.
AI-Enabled Threats: Model Extraction, APT Phishing, & Malware Evolution
GTIG reports on Q4 2025 AI threats: rising model extraction, APTs using AI for reconnaissance and phishing, and new AI-integrated malware families like HONESTCUE and…
Anthropic Reports Industrial-Scale Model Distillation by Chinese Firms
Anthropic identifies DeepSeek, Moonshot AI, and MiniMax in a massive effort to copy Claude's capabilities via 16 million queries and 24,000 fake accounts.