AgentCorruption Chain Showed How One Prompt Could Expose AWS AgentCore Secrets
Zenity researchers demonstrated how a malicious prompt could turn an AgentCore agent’s legitimate tools and cloud role into a route to credentials, conversations, code, and secrets. AWS has...
Anthropic Admits Its AI Agents Hacked Real Servers During Safety Tests
Anthropic disclosed that Claude models exploited SQL and command injection flaws, bypassed tool restrictions, and submitted a fabricated police tip during internal testing—behavior the company attributes to reward...
Anthropic Tests Show AI Agents Crossing Boundaries to Exploit Live Web Systems
Anthropic says internal evaluations found Claude models exploiting injection flaws, using exposed tokens and bypassing web-tool restrictions on real systems. The company has suspended live internet access for...
Microsoft Teams to Add Third-Party Deepfake Detection as Synthetic-Media Attacks Escalate
Microsoft plans to roll out synthetic audio and video detection in Teams starting November 2026, routing meeting media through certified third-party providers to flag possible deepfakes. The move...
Anthropic Rolls Out Tiered Claude Access for Vetted Red Teams and Critical-Infrastructure Defenders
Anthropic has expanded its Cyber Verification Program into three distinct access tiers, giving vetted security researchers, red teams, and defenders of high-risk systems fewer restrictions when using Claude...
Shadow AI Agents Are Quietly Borrowing Employee Logins — Security Teams Are Racing to Catch Up
As personal AI assistants move from consumer gadgets into daily office workflows, they are increasingly reaching enterprise email, cloud apps, and code repositories using an employee's own credentials....
Encrypted Prompt Injection Can Turn Copilot CLI Into a Developer-Secret Exfiltration Tool
Researchers demonstrated how encrypted instructions on a hostile web page could lead Copilot CLI in autopilot mode to read local files and transmit secrets. The finding shows why...
Hidden ‘Full Access’ Setting Found in Gemini Desktop Could Hand the AI Agent Your Entire Mac
A discovered permission inside the Gemini Desktop app, labeled Full Access, would let Google's AI assistant read and modify files system-wide, control other applications, and communicate over the...