NUWA RESEARCH RECORD
AgentPoison: Red-teaming LLM Agents via Poisoning Memory or Knowledge Bases
Original publication ↗ · BibTeX · RISThe reported experiments show that benign-task accuracy can coexist with triggered failures, so average ordinary-task utility does not alone establish retrieval integrity.
Research question
Can an agent remain useful on benign queries while retrieved memory steers it incorrectly?
Design & assumptions
The paper evaluates poisoned memory or knowledge bases across driving, question-answering and health-record agents, separating retrieval, action and end-to-end outcomes.
Findings
The reported experiments show that benign-task accuracy can coexist with triggered failures, so average ordinary-task utility does not alone establish retrieval integrity.
Limits & counterevidence
The threat model requires influence over stored retrieval material and specified triggers. Retrieval success, action selection and end-to-end success are distinct metrics, not interchangeable attack rates.
Source evidence
No structured numerical extraction is available in this reading package.
Reading scope & review
Selected original HTML sections and table context; not a complete paper/appendix audit or independent reproduction.
Model-assisted reading; independent evidence review pending.
Check the original source ↗