arXiv ScienceSearch

EXPLORE CONNECTIONS

LoginTrap: Uncovering Task-Agnostic Phishing-Style Indirect Prompt Injection Attacks against LLM-based Web Agents

Follow the relationships that help you find your next source.

Connections are not prepared for this record yet, or its source metadata has changed. The original record remains available while background snapshots are rebuilt.