Tarik A. Rashid
1 indexed paper
Recent (6 mo)
1With code
0Influential cites
0Benchmarked
0Publications per year
126
Top categories
Crypto×1NLP×1
Frequent co-authors
Research Timeline
2026
AgentShield: Deception-based Compromise Detection for Tool-using LLM Agents
AgentShield is a deception-based framework that detects successful indirect prompt injections in tool-using LLM agents across multiple languages by placing traps within the agent's tool interface.
Highlighted terms show continued research focus across papers