Tag: AI agents
Autonomous attacks run by DeepSeek agent on a single Telegram order
Unit 42 details a DeepSeek-driven agent that launched attacks after a single Telegram command.
Okta buys Permiso to pair identity posture with threat detection
Okta's deal for Permiso Security merges identity threat detection with posture management as AI agents multiply.
Cyera buys Oasis Security for $1B in identity-data merger
Data security company Cyera agreed to acquire agentic access management provider Oasis Security in a $1 billion deal.
Hugging face says autonomous AI agent breached its infrastructure
Hugging Face confirmed an autonomous AI agent breached its infrastructure through a malicious dataset in the company's data processing pipeline.
Memghost attack plants false memories in AI agents through a single email
Researchers discovered the MemGhost technique, which injects persistent false memories into AI agents via crafted email content, causing long-term manipulation…
AI coding assistants let researchers walk right past safety filters
Researchers achieve 100 percent jailbreak success against AI coding assistants including GitHub Copilot by decomposing harmful requests into ordinary coding…
Attackers Hide Commands in Web Code to Hijack AI Browsing Agents
Attackers are hiding malicious instructions in invisible HTML and structured data to trick AI browsing agents into making fraudulent payments…
AI Coding Agents Tricked Into Installing Malware From Clean Repos
Researchers at Mozilla's 0DIN platform show how AI coding agents can be tricked into executing malware by following standard setup…