SOURCE-LINKED INTELLIGENCE

ContextLeak: Exfiltrating LLM Agent Context via Malicious Tools

arXiv · AI, language, vision and robotics · article · Aug 28, 2026 · UTC

Exfiltrating an LLM agent's runtime context -- such as the user prompt, execution trajectory, and tool list -- poses severe security and privacy risks to users. Such attacks can be carried out via malicious tools and typically require three conditions: (1) the agent selects the malicious tool for task execution, (2) the agent passes its runtime context as input arguments to the tool, and (3) the tool's implementation transmits these inputs to an attacker-controlled endpoint. Existing work primarily focuses on conditions (1) and (3), leaving condition (2) largely unexplored, despite its critica

Read original source ↗ Open in workspace

recordType

paper

region

Global