Tag: LLM security
Browse our exclusive articles!
Inferring Communication Topology in LLM Multi-Agent Systems
Explore how Communication Inference Attack reveals topology in LLM-based multi-agent systems, exposing privacy risks and security vulnerabilities.
Preventing Policy Violations in LLM-Based AI Agents
Explore how PhantomPolicy and Sentinel frameworks help detect and prevent hidden policy violations in LLM-based agents for better compliance.
DeepGuard: Multi-Layer Secure Code Generation with LLMs
DeepGuard enhances secure code generation by aggregating multi-layer semantic cues in LLMs, improving vulnerability detection and preserving code correctne...
SkillSieve: Efficient Detection of Malicious AI Agent Skills
Discover SkillSieve, a fast, accurate triage framework detecting malicious AI agent skills with multi-layer analysis and LLM collaboration.
Prevent Data Leaks from Backdoored LLM Agents
Discover how backdoored LLM agents risk data leaks via tool calls and learn essential security measures to protect sensitive information effectively.