Attackers can inject indirect prompts in normal-looking repositories to trick Claude Code into spawning a reverse shell.

Decades-old Bash shell tricks can bypass safeguards in most open source AI coding agents, creating a new software supply ...

Makers of AI browsers make lofty promises. With a single prompt, users can ask one to find a restaurant in a particular part ...

Meta has restricted how its engineers use Anthropic's Claude Code and OpenAI's Codex, fearing it could accidentally distil a rival's model into its own.

Every prompt your team sends to a language model is a potential data-exfiltration event. According to Cyberhaven's 2026 AI ...