…Fixer Loops Published on Jun 8 Submitted by Ziqian Zhong on Jun 9 Carnegie Mellon University Authors: , , , , , Abstract Researchers identify widespread vulnerabilities in agent benchmark verification systems and develop an automated iterative…
…Can AI Agents Complete Real-World Post-Production Tasks? (2026) TerminalWorld: Benchmarking Agents on Real-World Terminal Tasks (2026) SWE-Marathon: Can Agents Autonomously Complete Ultra-Long-Horizon Software Work? (2026) Please…
… In contrast, SWE-Interact places agents in a realistic developer workflow: a carefully designed user simulator starts with vague or incomplete instructions, progressively reveals requirements, inspects the agent's workspace, and provides targeted feedback, revisions, and new constraints until the f… …
… 2026 WildClawBench: A Benchmark for Real-World, Long-Horizon Agent Evaluation 2026 AlphaEval: Evaluating Agents in Production 2026 RealClawBench: Live OpenClaw Benchmarks from Real Developer-Agent Sessions 2026 SWE-Marathon: Can Agents Autonomously Complete Ultra-Long-Horizon Software Work? …
…Can Frontier Models Solve Long-Horizon Auto Research and Engineering Tasks? Published on Jun 3 Submitted by taesiri on Jun 4 Authors: , , , Dongfu Jiang , , Hang Hua , , , , Lichi Li , , , , , , , , , Abstract AutoLab benchmark evaluates…
…Training Efficient Repository Explorer for Coding Agents Published on Jun 12 Submitted by Shaoqiu Zhang on Jun 16 Microsoft Authors: Shaoqiu Zhang , , Yuling Shi , , , , , Abstract FastContext separates repository exploration from code solving…
…Text World Models for LLM-based Agents Published on Jun 8 Submitted by Yixia Li on Jun 10 NLP Group in SUSTech Authors: Yixia Li , , , , , , , , , , , , , , , Abstract Text world models serve as transition…
… We build an extension for Claude Code, enabling developers to monitor and improve their own agentic systems. …
…Toward Trustworthy Evaluation of Autonomous Agents (2026) Toward Scalable Terminal Task Synthesis via Skill Graphs (2026) EnterpriseLab: A Full-Stack Platform for developing and deploying agents in Enterprises (2026) Please give a…
…Position-Aware Undetectable Skill Injection on LLM Agents Published on Jun 6 Submitted by Haochang Hao on Jun 11 University of Illinois Chicago Authors: , , , , , , Abstract POISE is a stealthy skill-poisoning attack…