…Improving LLM-Authored Agent Skills via Trace-Conditioned Skill Revision (2026) SkillOpt: Executive Strategy for Self-Evolving Agent Skills (2026) CODESKILL: Learning Self-Evolving Skills for Coding Agents (2026) Socratic-SWE: Self…
…Online Adaptation for Self-Improving Foundation Agents Published on May 11 Submitted by Seth Karten on May 13 Princeton University Authors: Seth Karten , , , , , , , Abstract A self-improving AI system for embodied agents…
…errors can propagate across agents and interaction rounds, producing failures that are difficult to diagnose and rarely translate into structural self-improvement . Existing surveys cover individual agent capabilities, multi- agent collaboration , or…
…Scaling Computer-Use Agents via Temporal Visual Redundancy Reduction Published on Jun 5 Submitted by Amirhossein Abaskohi on Jun 11 Microsoft Research Authors: , , , , , Abstract ReVision improves computer-use agent efficiency by removing…
… The following papers were recommended by the Semantic Scholar API The Blind Spot of Agent Safety: How Benign User Instructions Expose Critical Vulnerabilities in Computer-Use Agents 2026 Plant, Persist, Trigger: Sleeper Attack on Large Language Model Agents 2026 AIRGuard: Guarding Agent Actions wit… …
Papers arxiv:2606.01770 Adaptive Auto-Harness: Sustained Self-Improvement for Agentic System Deployment on Open-Ended Task Streams Published on Jun 1 Submitted by Zewen Liu on Jun 3 Authors: Zewen Liu , , , , , , , , , Abstract Adaptive Auto-Harness framework addresses dynamic task streams by decom… …
…Multi-Agent Data Generation for Accurate and Generalizable Function-Calling (2026) From Raw Experience to Skill Consumption: A Systematic Study of Model-Generated Agent Skills (2026) Pioneer Agent: Continual Improvement of Small…
Papers arxiv:2606.06416 Unsupervised Skill Discovery for Agentic Data Analysis Published on Jun 4 Submitted by Shumin Deng on Jun 5 ZJUNLP Authors: , , , Shuofei Qiao , , , Abstract DataCOPE is an unsupervised framework that discovers reusable data-analysis skills through verifier-guided exploratio… …
… Ablation studies show that directly prompting agents with raw guides can degrade performance, while both structured skill construction and trajectory-driven revision are necessary for the observed improvements. …
… SkillAdaptor improves over no-skill and skill-adaptation baselines on all three suites, with the largest single-metric improvements of +1.5 points on PinchBench Avg Score%, +1.8 on Claw-Eval Avg Score, and +1.7 on WebShop success rate. …
