AI RESEARCH
MemMorph: Tool Hijacking in LLM Agents via Memory Poisoning
arXiv CS.AI
•
ArXi:2605.26154v1 Announce Type: cross LLM-driven agents are capable of selecting external tools to complete users' tasks. However, attackers could compromise such process, steering agents toward inappropriate/wrong tools and enabling malicious actions. Most existing attacks primarily manipulate the tool metadata, which is easily detectable by auditing and may lose effectiveness as modern agents increasingly adopt memory modules to refine tool selection policies through accumulated experience.