Attacks on AI Research AI security importance 60 1 min read In your stack
Weighing the Envelope: Component Weight and Arbitration in Envelope-Layer Defenses for a Single LLM Agent
Researchers measure how pre-planted malicious MCP tools and poisoned long-term memory influence tool-invocation decisions in a single LLM agent, without attacker access to the conversation.
ReviewReview agent tool sources and long-term memory for untrusted or pre-planted content
Techniques
Paper
Sources
arXiv (AI security papers) primary
One-line summary and action are AI-assisted. Read the source before acting.