Agent Skills can specify procedural and resource obligations for tool use, and language models instantiate them as concrete programs. However, when models turn this guidance into code for existing too...
ClawSentry is an open‑source, framework‑agnostic security supervision gateway designed to protect autonomous large language model (LLM) agents from progressive risks that can arise at four points in the agent control loop: skill admission, invocation‑time intent, execution‑time effect, and post‑action consequence. It introduces a multi‑tier decision engine—deterministic L1, rule‑anchored L2, and read‑only L3—alongside a First‑Use Skill Package Review (FSPR) and an Agent Harness Protocol (AHP) that applies a single policy across multiple agent runtimes without modifying their internals. Evaluation on SkillInject and the SkillsSafety benchmark shows that ClawSentry significantly reduces contextual adversarial skill risk (ASR) while maintaining high task success rates (TSR).
By Kai Wang, Zeming Wei, BiaoJie Zeng, Chang Jin, An Wang, Xiaokun Luan, Zhixiao Lin, Jingjing Qu, Xia Hu, Xingcheng Xu
The paper introduces JAZ, a minimalist LLM agent framework that centers on a single primitive called “invoke”, which allows an LLM to write and execute arbitrary code, including recursive calls, while treating all inputs and interaction history as variables in the code environment. JAZ provides built‑in hooks for constraints and monitoring but relies solely on prompting, without external tools, memory systems, or file‑system access. Experiments show that JAZ “invoke” outperforms specialized external harnesses such as Letta (MemGPT) and ACE on long‑horizon recall tasks and continual self‑improvement, achieving higher accuracy at lower cost.
By Zhening Li, Joshua Liu, Mateja Vukelic, Nicole Shen, Supriya Lall, Amitayush Thakur, Alex Zhang, Omar Khattab, Jonathan Light, Armando Solar-Lezama
ContrAgent is a contract‑based framework that provides symbolic temporal supervision for large language model agents. It records an agent’s tool‑call sequence as a trace of checkable predicates and formalizes desired behaviors with assume‑guarantee contracts expressed in linear temporal logic over finite traces (LTLf). Each contract is compiled into a deterministic finite automaton that both gates actions online and evaluates recorded traces offline, enabling deterministic, reproducible verdicts and significantly lower per‑call latency compared to existing LLM‑judge and rule‑based guardrail baselines.
By Yifeng Xiao, Pierluigi Nuzzo
The paper introduces Many-Tier Instruction Hierarchy (ManyIH), a new framework for resolving conflicts among instructions with arbitrarily many privilege levels in large language model agents. It presents ManyIH-Bench, a benchmark featuring 853 agentic tasks that require navigating up to 12 levels of conflicting instructions across 46 real-world agents. Experiments show current models achieve only about 40% accuracy when instruction conflict scales, highlighting a gap in fine-grained, scalable conflict resolution.
By Jingyu Zhang, Tianjian Li, William Jurayj, Hongyuan Zhan, Benjamin Van Durme, Daniel Khashabi
arXiv:2606. 24311v1 Announce Type: new Abstract: As large language model (LLM) agents are applied to longer tasks, they increasingly modify workspace state across multiple rounds of iteration.
By Kailong Ren, Fubo Sun, Jiachen Liu, Liu Yang, Zimo Yin, Jiaying Li, Congli Yin, Ming He, Yu Huo, Jiawei Liu, Zeping Chen, Yubin Huangfu, Ronghua Li, Yixuan Wu, Xing Su, Yanzhi Xu, Likang Wu, Hongke Zhao, Lei Zhang, Xiaohui Geng, Jianping Fan