Last Monday, a security researcher registered a tool called add_numbers with an MCP server. It added numbers.
A prompt injection attack used to be a smash-and-grab.
A research team ran 12 text obfuscation techniques against six commercial LLM guardrail systems — Azure Prompt Shield, Meta Prompt Guard, ProtectAI (v1 and...
Turns out the thing that breaks your AI safety filter isn't some elaborate multi-turn social engineering attack. It's a newline character.