OpenAI pauses its "most capable models" after agents exploit loopholes and leak data
CVE-2026-82385: Apache Roller: Weblog template include escapes the Velocity sandbox and reads classpath files
Revealing the details of how OpenAI agents hacked Hugging Face
in the wild 15 min
Hard Stop: Kernel-Level Preemption and Containment for Rogue Agentic Execution
Worries About an AI Internet Takeover Gain New Urgency Among Doomsday Scenarios
in the wild 5 min
xdg-dbus-proxy 0.1.9 fixes sandbox escape CVE-2026-94422
Re: bubblewrap 0.12.0 fixes writes outside sandbox
After spending billions, OpenAI still has gaps in its cybersecurity
4 min
Researchers escape OpenAI Codex sandbox to run commands on host
2 sources 4 min
Risky Bulletin: Gemini hacked three companies too
15 min1
OpenAI and Anthropic oversold AI security breaches
in the wild 4 min
Google's Gemini also accidentally hacked three real companies during security testing
2 min
Google Gemini AI Hacked 3 Real Companies during a Cybersecurity Test
4 min2
Gemini hacked three companies in first known breakout by Google's AI
in the wild
Inference-Engine Fingerprinting Attacks are Practical: Exploring Model-Driven Environmental Discovery, Exploitation, and Escape
USN-8779-1: Bubblewrap vulnerabilities
Critical Docker Sandboxes Flaw Lets Malicious Guest Code Read and Modify macOS Host Files
4 min
Critical Docker Sandbox Vulnerabilities Enable Malicious Guests to Escape Isolated microVM Workspaces
3 min
AI labs want in-house auditors — but maybe they should shut the front door first
6 min
Apple Releases iOS 27 Security Update to Fix Over 120 Vulnerabilities
3 min
CVE-2026-77147: Apache Syncope: Groovy Sandbox escape for empty CommandArgs
Anthropic CEO says AI swarm could 'take over the Internet' in 6-12 months
9 min3
Anthropic reveals rogue AI agents hate CAPTCHAs, just like you
5 min1