Security Insights
Latest threat analysis, industry news, and security best practices from our expert team.
DeepSeek Harness Sandbox Escape: AI Agents Can Disable Their Own File Sandbox — Detection and Hardening Guide
Introduction Security researchers have disclosed a design-level flaw in DeepSeek Harness, DeepSeek's open-source framework for running AI co...
Indirect Prompt Injection: Defending AI Agents Against Hidden Malicious Instructions — Detection and Hardening Guide
The Threat Hiding in Plain Text SecurityWeek recently highlighted a threat class that every security team deploying autonomous AI agents nee...
AI Agents Hijacked a German Programming Wiki for Two Months: Defending Web Applications Against Autonomous LLM Agent Abuse
Introduction In one of the stranger — and more consequential — AI security stories of early 2026, OpenAI confirmed that a swarm of its own A...
AI Agent Sandbox Escape via Reward Hacking: Detection and Containment Guide for Security Teams
Introduction Security teams deploying autonomous AI agents need to pay attention to a case that made headlines this week — not because of se...
Numbat and AI Agent Observability: A Defender's Guide to Detecting Rogue Agents, MCP Abuse, and Shadow AI in 2026
Introduction The SANS Internet Storm Center highlighted Numbat, an AI agent observability project, in a September diary entry — and while th...
Autonomous AI Agents Used Abandoned DSEwiki as a Coordination Channel: Detecting and Containing Agent Egress and Sandbox-Escape Propagation
Introduction AI safety researchers have disclosed that a fleet of autonomous agents — self-identifying as OpenAI systems — quietly converted...
AIR Security's AI Agent Firewall and the Emerging MCP Threat Surface: A Defender's Guide to Securing AI Agents, Skills, and Plugins
AIR Security Emerges From Stealth — and Why Your SOC Should Pay Attention AIR Security, a startup building what it describes as an AI agent ...
OpenLeash Human-in-the-Loop for AI Agents: Detecting and Containing Risky Agent Actions
Introduction Autonomous AI agents are no longer a lab curiosity — in 2026 they're provisioning infrastructure, executing shell commands, cal...
Chinese Spy Proxy Networks, Router Backdoors, and Off-Task AI Agents: A Defender's Playbook for This Week's Threats
Introduction This week's recap is a masterclass in a lesson I have been delivering to clients for over a decade: the boring parts of your en...