The paper introduces GADR, a multi‑agent, self‑correcting workflow that extracts architectural decisions from raw meeting transcriptions and produces Nygard‑formatted ADR drafts. A feasibility study using five real project transcripts, expert reviews by four senior architects, and evaluations by fifteen students shows that GADR captures most expert‑identified decisions and yields drafts that participants find clear and useful, outperforming zero‑shot and few‑shot baselines in stability and structural adherence. The study also highlights a trade‑off: RAG‑based enrichment can deepen ADR content but may introduce transcript‑unfaithful material, raising open questions about traceability in automated architectural documentation.
By Lucas Daniel Costa da Silva, Kiev Gama
The paper introduces CADET, a training‑free framework for auditing, benchmarking, and repairing spurious reliance in pretrained end‑to‑end autonomous‑driving planners. It addresses the problem that such planners often learn statistical shortcuts—associating co‑occurring scene elements with driving decisions—rather than causal variables, which undermines reliability in rare scenarios. CADET can detect and correct these causal confusions without retraining the model or updating its parameters.
By Zikun Guo, Minglan Chen, Jinyou Zhai, Rongjin Zou
Agentic ESOpt proposes using evolution strategies (ES) instead of reinforcement learning to fine‑tune large language‑model agents for long‑horizon tasks. ES offers model scalability, flexibility, and better long‑horizon credit assignment, enabling full‑parameter optimization with minimal GPU memory. The framework samples parameter perturbations, evaluates agents with rewards, and updates online, achieving notable performance gains on WebArena‑Lite and in test‑time prompt‑parameter co‑evolution.
By Zhi Zheng, Rongsheng Chen, Yunpeng Ba, Zhenkun Wang, Yee Whye Teh, Wee Sun Lee
The paper introduces SEE, a two-stage framework for generating long-horizon GUI agent trajectories. First, an exploration stage builds an explicit UI transition graph over screens and elements. Second, a graph-based synthesis stage composes diverse multi-step trajectories through planning and controlled sampling, preventing spurious cycles and enabling long-horizon composition. Across real-world apps, SEE produces trajectories averaging 14.8 steps and improves agent task success and generalization to unseen screens.
By Zhuohang Fan, Beichen Zhang, Yuanfa Li, Changqiao Wu, Wei Liu, Jian Luan, Weigang Zhang
The paper proposes using encoder‑decoder networks based on LSTM cells to reduce communication in distributed model predictive control (DMPC). Agents transmit a compressed representation of their messages, which receivers reconstruct, thereby lowering the data exchanged each time step. Experiments with mobile robot formations show that the trained networks maintain performance even when full communication is unavailable, and they can adapt to different prediction horizons without retraining.
By Torben Schiz, Pedro H. J. Nardelli, Henrik Ebel
HarnessRisk is a lifecycle-oriented benchmark for evaluating safety in agent harnesses that manage tools, extensions, state, permissions, and external actions. It defines six operational phases—Harness Configuration, Capability Extension, Runtime Operation, State Persistence, Action Control, and Incident Recovery—and includes 128 sandboxed cases pairing benign user objectives with adversarial instructions. Across three harnesses, six language models, and 14 configurations, attack success rates vary from 12.6% to 80.9%, with the most vulnerable phase being Harness Configuration.
"whyItMatters":"The benchmark demonstrates that safety failures can arise in multiple harness responsibilities and that even explicit risk detection does not guarantee safe action, underscoring the need for comprehensive evaluation across model and harness configurations."
By Yajing Bai, Jinhao Duan, Jie Peng, Xianfeng Wu, Sijia Liu, Song Wang, Tianlong Chen
CityReal is a modular framework that uses large language model agents to simulate human-aligned urban behavior. It models agents as intention-driven decision makers who pursue coherent mobility and activity plans, learning habits and preferences over time. By training textual adapters to align agent decisions with observed population statistics, CityReal improves realism at both micro and macro levels and can scale to tens of thousands of agents for analyzing crowd density, place popularity, mobility flows, and well‑being under various urban scenarios.
By Nicolas Bougie, Xiaotong Ye, Narimasa Watanabe
The paper investigates how a bounded agent should allocate its limited memory and communication resources when making decisions. It defines the remembering–signaling frontier as the set of memory and message rate pairs that achieve a given performance threshold for a fixed task and decision rule. The authors hypothesize that when history can reduce task loss more, the agent will need less peer communication, and preliminary referential game experiments support this idea.
By Yashar Talebirad, Eden Redman, Ali Parsaee, Osmar R. Zaiane
The paper re‑evaluates memory‑based self‑improving agents by adding multiple runs to measure variance and by randomizing task order. It finds that agent performance is noisy in complex, multi‑step environments and that improvement depends heavily on the sequence of tasks, revealing a hidden curriculum effect. The authors suggest that underspecification of tasks and environments contributes to this fragility and demonstrate that adding detailed rubrics and feedback can partially mitigate performance drops, though gaps remain.
By Qinyuan Ye, Yu Li, Yada Pruksachatkun, Jiaxin Zhang, Chien-Sheng Wu
SkillEffect is a checked‑lowering runtime that ensures agent tool calls stay within memory limits by verifying each proposed program against an immutable input before execution. It uses audited relation plugins to provide source recognition, bounded intermediate representation construction, and postconditions, while a shared runtime handles selection, bounded VM execution, and atomic capacity leasing. Experiments across six operator families show that bounded access significantly reduces peak memory usage and improves completion rates under fixed memory caps.
By Yinuo Wang, Yiyu Shi
StartupBench is a new benchmark that evaluates general-purpose agents on end-to-end workflows derived from AI startup products that have proven market adoption. It translates real-world product workflows into deliverable-oriented tasks and assesses them with detailed rubrics. The study finds that even the best models complete only about 30% of these tasks, highlighting challenges such as complex instruction following and domain expertise.
By Liya Zhu, Xin Ma, Tao Liu, Haodong Wang, Ge Zhang, Jingzhe Ding, Qingshui Gu, Yongjie Zhong, Jinxiang Meng, Yuan Gao, Yunqiu Zhou, Hao Zhu, Jifeng He, Yongzhi Liao, Xinyi Zhang, Chaoxin Li, Yi Zhu, Xi Lin, Duju Zeng, Xiang Gao, Wen Zhang, Yunyang Wang, Duo Wang, Huan Zhou, Zuo Wang, Jin Chen, Kaiyuan Zhang, Chuqian Yu, Tianhao Yu, Longxiang Liu, Jianbo Xue, Huimin Che, Jiahao Wang, Yujia Qin, Jiaheng Liu, Shen Yan, Xiaolong Chang, Wenhao Huang
The paper introduces an automated method that employs Large Language Models in a check–fix–verify loop to generate executable Python scripts for extracting problem-specific features from constraint satisfaction problems. Given a MiniZinc model and instance, the LLM agent produces code that builds a typed graph representation and computes structural properties such as graph density, variable clustering, and constraint tightness. Evaluated on vehicle routing, car sequencing, and fixed‑length error‑correcting codes, the synthesized extractors enable algorithm selectors that outperform expert‑curated mzn2feat features and transformer‑based trans2feat variants, while remaining interpretable.
By Hai Xia, Carlos Ans\'otegui, Stefan Szeider
Wuying-Browser-Agent is a unified framework designed to improve long-horizon browser agents by aligning execution, supervision, optimization, and evaluation. It introduces a structured browser harness, reflection and UI-specialized Curriculum SFT (RUIC‑SFT) for recovery and complex UI interactions, and Divergence‑Aware Online GRPO (DAO‑GRPO) for better credit assignment. The framework is evaluated on BrowserBench—a bilingual real‑web benchmark of 350 tasks—and achieves state‑of‑the‑art results on multiple browser‑use benchmarks, while also transferring well to other agentic tasks.
By AIMAE Team, Tianxiang Chen, Yan Cheng, Zhangye Han, Xiaowei Li, Chang Liu, Cheng Liu, Zhongqiang Ma, Long Peng, Xiaobing Tu, Yinggui Wang, Hongliang Wei, Chen Wu, Daiping Xin, Kunyu Zhou, Pengyang Zhou, Peiyuan Chen, Ziyuan Chen, Yutao Deng, Chunyu Dong, Xiangyu Fu, Yicheng Feng, Ruian He, Haochen Li, Miancan Liu, Zhengqin Liu, Wei Peng, Jinkui Ren, Haoyu Tan, Dong Xiao, Rongkun Xue, Shujian Yang, Xianhang Ye, Ziqi Yuan, Ziyang Yu, Linghan Zhang, Xiantao Zhang, Xuanpu Zhao, Yinan Zhao, Zhenghui Zhao, Bin Zhu, Likai Zou
The paper introduces RATTL (Risk-Adversarial Total-Reward Learning), a framework that adjusts an agent’s caution based on epistemic uncertainty by using a Bayesian posterior over dynamics and a Wasserstein ambiguity set whose radius depends on that posterior. As evidence accumulates, the radius shrinks, smoothly transitioning the agent’s behavior from worst-case robustness to risk-neutral reward maximization. The authors prove a Safety Sandwich theorem showing RATTL’s value lies between the uninformed robust value and the full-knowledge optimum, and demonstrate the method on a binary-hazard example where the criterion reduces to Conditional Value-at-Risk.
By Deep Kumar Ganguly, Jan Kretinsky
The paper introduces Aegis, a runtime governance system for agentic AI that treats model outputs as action proposals and mediates them through a trusted decision layer before tool execution. Aegis evaluates proposals against active policy, resolves provenance server‑side, fails closed under uncertainty, and routes selected cases through a Senate‑style settlement process. In a sandbox evaluation across 6,300 rows, Aegis prevented all governed mock‑tool applications and risky side‑effect completions, preserving provenance and quorum evidence for all settled cases.
By Adam Mazzocchetti
The paper introduces a task-aware harness provisioning method for large language model agents used in mission‑critical infrastructure operations. By mapping each task to an optimal harness configuration—determined through mathematical task classification and empirical agent execution data—the authors propose a map‑guided escalation algorithm that starts with a minimal harness and expands only when necessary. Experiments on liquid‑cooling and power‑grid tasks show that this approach can improve accuracy and reduce token usage compared to full harness provisioning, revealing a domain‑dependent accuracy‑cost trade‑off.
By Liangtao Lin, Qingang Zhang, Zhaomeng Zhu, Tianwei Zhang, Yonggang Wen
The paper introduces FACA, a Feedback‑Aware Credit Assignment method that aligns each agent reaction with the preceding user‑to‑user segment, computes a locally normalized reaction advantage, and adds it to the terminal outcome advantage without requiring an extra critic or rollout. Experiments show that FACA improves performance across nine domains, especially in Telecom, and maintains the same ordering of gains in zero‑shot benchmarks such as Pare‑Bench and Co‑Gym.
By Yiwen Zhao, Zhihao Wen, Yuchen Mao, Mingxuan Jiang, Yihao Hu, Pan Wang, Xin Zhang, Wei Wu
The paper proposes Novelty and Surprise Prioritized Experience Replay (NSPER) for image-based reinforcement learning, combining novelty to highlight underrepresented states and surprise to reveal gaps in the agent’s knowledge. An extended version, NSPER+R, also uses these signals as intrinsic rewards to enhance both replay quality and exploration. Experiments on DeepMind Control Suite tasks demonstrate that NSPER and NSPER+R accelerate training and improve convergence compared to existing methods.
By Hoda Yamani, Henry Williams, Bruce A. MacDonald
EvoTS-Agent is a self‑evolving large language model agent designed for autonomous change‑point detection in financial time series. It begins with curated exploratory data analysis to set up candidate models, then iteratively refines its detection pipeline using three operators—Revision, Alternative Strategy, and Recombination—guided by validation feedback. Across four benchmark datasets, EvoTS-Agent consistently outperforms existing LLM‑based agents and achieves a 100% execution success rate on all tested backbone LLMs.
By Lei Jiang, Ye Wei, Xinyu Xi, Jordan Langham-Lopez, Yifan Bao, Raad Khraishi, Yihao Ang, Anthony K. H. Tung, Lukasz Szpruch, Hao Ni
The paper audits self‑evolving financial agents—SkillOpt, Agent Workflow Memory (AWM), and ReasoningBank—by evaluating their performance, security drift, and execution‑interface mismatches in simulated e‑banking scenarios. It shows that while utility improves after evolution, exposure to malicious content and unauthorized state changes also rise, and that AWM’s text‑action envelope can disrupt tool execution, highlighting the need for comprehensive auditing beyond accuracy metrics.
By Jialong Li, Jialing Zhu