Concept工具投毒
Tool Poisoning
An attack method in MCP/Agent scenarios in which an attacker embeds hidden instructions in tool descriptions, schemas or default values, exploiting the protocol's trust assumptions about tool metadata so that malicious instructions are folded into the system context and executed.
Timeline (last 90 days)
Sep 28
The attack in which a server silently changes tool definitions after being trusted is referred to by researchers as Tool Poisoning.
Unverified50%