← Back ◬ AI & Machine Learning Mar 25, 2026

T-MAP: Red-Teaming LLM Agents with Trajectory-aware Evolutionary Search

arXiv Security Archived Mar 25, 2026 ✓ Full text saved

arXiv:2603.22341v1 Announce Type: new Abstract: While prior red-teaming efforts have focused on eliciting harmful text outputs from large language models (LLMs), such approaches fail to capture agent-specific vulnerabilities that emerge through multi-step tool execution, particularly in rapidly growing ecosystems such as the Model Context Protocol (MCP). To address this gap, we propose a trajectory-aware evolutionary search method, T-MAP, which leverages execution trajectories to guide the disco

Full text archived locally

✦ AI Summary · Claude Sonnet

Computer Science > Cryptography and Security [Submitted on 21 Mar 2026] T-MAP: Red-Teaming LLM Agents with Trajectory-aware Evolutionary Search Hyomin Lee, Sangwoo Park, Yumin Choi, Sohyun An, Seanie Lee, Sung Ju Hwang While prior red-teaming efforts have focused on eliciting harmful text outputs from large language models (LLMs), such approaches fail to capture agent-specific vulnerabilities that emerge through multi-step tool execution, particularly in rapidly growing ecosystems such as the Model Context Protocol (MCP). To address this gap, we propose a trajectory-aware evolutionary search method, T-MAP, which leverages execution trajectories to guide the discovery of adversarial prompts. Our approach enables the automatic generation of attacks that not only bypass safety guardrails but also reliably realize harmful objectives through actual tool interactions. Empirical evaluations across diverse MCP environments demonstrate that T-MAP substantially outperforms baselines in attack realization rate (ARR) and remains effective against frontier models, including GPT-5.2, Gemini-3-Pro, Qwen3.5, and GLM-5, thereby revealing previously underexplored vulnerabilities in autonomous LLM agents. Subjects: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI); Computation and Language (cs.CL) Cite as: arXiv:2603.22341 [cs.CR] (or arXiv:2603.22341v1 [cs.CR] for this version) https://doi.org/10.48550/arXiv.2603.22341 Focus to learn more Submission history From: Hyomin Lee [view email] [v1] Sat, 21 Mar 2026 12:33:34 UTC (726 KB) Access Paper: HTML (experimental) view license Current browse context: cs.CR < prev | next > new | recent | 2026-03 Change to browse by: cs cs.AI cs.CL References & Citations NASA ADS Google Scholar Semantic Scholar Export BibTeX Citation Bookmark Bibliographic Tools Bibliographic and Citation Tools Bibliographic Explorer Toggle Bibliographic Explorer (What is the Explorer?) Connected Papers Toggle Connected Papers (What is Connected Papers?) Litmaps Toggle Litmaps (What is Litmaps?) scite.ai Toggle scite Smart Citations (What are Smart Citations?) Code, Data, Media Demos Related Papers About arXivLabs Which authors of this paper are endorsers? | Disable MathJax (What is MathJax?)

💬 Team Notes