Tool Poisoning Attack
Named by 2 kept items in this workspace.
Search every workspace for Tool Poisoning Attack. It matches the name as written, so it can find other things that share it.
The items that name it 2
Exploits LLM-server interactions to inject malicious prompts via MCP
Malicious instructions embedded within a tool's metadata (not executed) to manipulate LLM agent behavior; exploits agents' instruction-following ability rather than tool output injection
All techniques in this workspace.