AI news archive.
4,966 articles filed under Agents, newest first, from the outlets listed on the sources page.
- arXiv cs.AIagentsFalse Frontiers: Diagnosing and Mitigating Co-Cheating in Self-Evolving Search Agents
arXiv:2609.39102v2 Announce Type: replace-cross Abstract: Self-evolving search agents build their own training curricula by jointly optimizing a proposer that generates questions and a solver that answers them. This clos…
- arXiv cs.AIagentsDAGent: Evaluate-then-Grow Planning for Deep Research Agents
arXiv:2609.39154v1 Announce Type: cross Abstract: Deep research tasks require agents to navigate large knowledge spaces, synthesize evidence across many sources, and adapt their plans as findings emerge. Directed acyclic…
- arXiv cs.AIagentsScale and Selection: What Makes Automatic Harness Evolution Work for Visual-Interface Robot Agents
arXiv:2609.39304v4 Announce Type: replace-cross Abstract: When an off-the-shelf coding agent is used directly as a robot policy, observing a browser-based 3D interface through screenshots and acting by posing a virtual t…
- arXiv cs.AIagentsReSAIL: Mitigating Collapse in Iterative Agent Self-Distillation
arXiv:2609.39306v1 Announce Type: cross Abstract: Iterative self-distillation enables LLM agents to learn from successive deployments, offering a path toward recursive self-improvement (RSI). Yet our experiments with exi…
- arXiv cs.AIagentsNarrativeSteward: Coordinating Delegation, Guidance, and Verification in Agent-Assisted Interactive Narrative Authoring
arXiv:2609.39333v1 Announce Type: cross Abstract: Autonomous AI agents can turn authors' goals into interactive narratives by independently organizing and carrying out generation and revision. As agents generate and revi…
- arXiv cs.AIagentsBelief-Based Maximum Occupancy Principle and Active Inference
arXiv:2609.39342v1 Announce Type: cross Abstract: Intrinsic motivation plays a central role in adaptive and goal-directed behavior by conferring agents reward-independent objectives and biases useful to act in noisy and…
- arXiv cs.AIagentsHiding in Plain Sight: Decoupling Pretext from Actuation for Skill Poisoning in LLM Agents
arXiv:2609.39352v1 Announce Type: cross Abstract: LLM agents increasingly rely on reusable Skills for complex, multi-step tasks, creating a critical supply-chain attack surface where poisoned Skill content steers agent d…
- arXiv cs.AIagentsFrom Imitation to Reward Discovery: On-Policy Warmup for Agentic RL
arXiv:2609.39436v1 Announce Type: cross Abstract: Reinforcement learning with a verifiable reward (RLVR) offers a scalable approach to training language-model agents, yet sparse outcome rewards can leave early training w…
- arXiv cs.AIagentsActionGuard: Tool Call Authorization under Poisoned Skills
arXiv:2609.39450v1 Announce Type: cross Abstract: LLM-based agents extend their capabilities through third-party skills that provide task-specific instructions, scripts, and tool-use procedures. However, malicious instru…
- arXiv cs.AIagentsSpeculative Safety Honeypot: Toward Proactive Defense Against Multi-turn Agent Attacks
arXiv:2609.39549v1 Announce Type: cross Abstract: As Large Language Model (LLM) agents are increasingly deployed in complex environments, multi-turn interaction attacks have become a significant security challenge. Exist…
- arXiv cs.AIagentsPretext: Defeating Malicious Skill Detection Frameworks for AI Agents
arXiv:2609.39607v2 Announce Type: replace-cross Abstract: Skills extend an agent's capabilities by injecting instructions and information into the context, and are widely used by agents such as OpenClaw and Claude Code.…
- arXiv cs.AIagentsSEPAL: Separated Expert Pairs with Answer-Level Fusion for Reliable LLM Collaboration
arXiv:2609.39645v1 Announce Type: cross Abstract: Multi-agent collaboration lets large language models (LLMs) improve question answering through deliberation and feedback. Yet shared discussion couples correction with ex…
- arXiv cs.AIagentsDoGBench: Can Agents Meet Expert Standards for User-Facing Documentation?
arXiv:2609.39909v1 Announce Type: cross Abstract: We introduce DoGBENCH (Documentation Generation Benchmark), to our knowledge, the first benchmark for generating and maintaining real user-facing software documentation.…
- arXiv cs.AIagentsTRACE: Trajectory Selection for Parallel Scaling of Search Agents
arXiv:2609.39912v1 Announce Type: cross Abstract: Parallel search may generate a correct answer that final-answer voting fails to select. We formulate this consolidation stage as trajectory selection and introduce TRACE…
- arXiv cs.AIagentsLearning When and How to Intervene: A Hindsight-Distilled Sentinel for Coding Agents
arXiv:2609.39957v1 Announce Type: cross Abstract: Coding agents solve repository-level tasks through sequences of actions, where a single erroneous action can misdirect subsequent decisions and increase recovery costs. E…
- arXiv cs.AIagentsMid-Harness: Scaling Actions Between Model and Harness for Terminal Agents
arXiv:2609.39982v1 Announce Type: cross Abstract: Terminal agents act through stochastic model generations, yet the ability to generate a useful action does not ensure its reliable execution. A poor command (e.g., wrong…
- arXiv cs.AIagentsGame-Guided Skill Discovery through Self-Play for Playable Agent Control
arXiv:2609.40137v1 Announce Type: cross Abstract: We present Game-Guided Skill Discovery (GGSD), a framework that uses self-play in games to discover motor skills that are directly playable by humans. Playable skills pro…
- arXiv cs.AIagentsMemLife: Curating and Reasoning over Long-Term Egocentric Video Memories
arXiv:2609.40195v1 Announce Type: cross Abstract: Long-term egocentric video enables personalized AI assistants to reason about daily life. However, as video histories grow to hundreds of hours spanning months or years,…
- arXiv cs.AIagentsPhantomEnvironments: Training LLM Agents in Fictional Worlds
arXiv:2609.40221v1 Announce Type: cross Abstract: Training LLM agents with reinforcement learning (RL) is bottlenecked by environments, which must provide verifiable rewards, support long-horizon interaction, and scale c…
- arXiv cs.AIagentsEviRover: Reinforcing Agentic Perception Beyond a Glance
arXiv:2609.40230v1 Announce Type: cross Abstract: Visual perception is conventionally formulated as a one-shot prediction from a single glance at the image, under the assumption that the image content and the model's par…
- arXiv cs.AIagentsComputerSD: Online Self-Distillation from Real-Time Feedback for Computer-Use Agents
arXiv:2609.40253v2 Announce Type: replace-cross Abstract: Online training enables computer-use agents (CUAs) to improve through interaction with executable environments. However, existing methods primarily rely on sparse…
- arXiv cs.AIagentscua-speedrun: Standardized Benchmarking of the Speed of Computer-Use Agents
arXiv:2609.40284v1 Announce Type: cross Abstract: Computer use agents (CUAs), which use graphical user interfaces (GUIs) to complete tasks on a computer, have recently surpassed human performance on many standard benchma…
- arXiv cs.AIagentsDynaHarness: A Dynamic Physical Harness for Self-Evolving Robot Agents
arXiv:2609.40306v1 Announce Type: cross Abstract: Pretrained robot policies provide useful action priors, but long-horizon manipulation still requires coordination between semantic reasoning and physical execution. Seman…
- arXiv cs.AIagentsA memory-based active inference model of DishBrain-like adaptive behaviour
arXiv:2508.06980v2 Announce Type: replace Abstract: Recent and rapid advances in artificial intelligence (AI) make it increasingly important to understand the foundations of adaptive behaviour in autonomous agents, espec…
- arXiv cs.AIagentsCo-Evolving Agents: Learning from Failures as Hard Negatives
arXiv:2511.22254v5 Announce Type: replace Abstract: Self-evolving agents improve their performance on long-horizon tasks by learning from their own interactions with an environment. A common approach uses the resulting f…
- arXiv cs.AIagentsEvaluating Agents Across Runtime Contracts: When Mismatch Costs Efficiency or Quality
arXiv:2603.01209v3 Announce Type: replace Abstract: In CodeAct, language-model agents write Python that calls tools and use execution feedback to choose actions. Persistent runtimes preserve Python variables between acti…
- arXiv cs.AIagentsSkillMaster: Toward Autonomous Skill Mastery in LLM Agents
arXiv:2605.08693v3 Announce Type: replace Abstract: Skills provide an effective mechanism for improving LLM agents on complex tasks, yet in existing agent frameworks, their creation, refinement, and selection are typical…
- arXiv cs.AIagentsGeneralizing the Turing Test to Interactive Agents
arXiv:2605.10851v2 Announce Type: replace Abstract: We initiate the study of the Generalized Turing Test (GTT), a formal generalization of Turing's imitation game from humans to arbitrary interactive agents. For agents $…
- arXiv cs.AIagentsASH: Agents that Self-Hone in Long-Horizon Worlds
arXiv:2605.14211v5 Announce Type: replace Abstract: Long-horizon visuomotor tasks remain a fundamental challenge in AI, as current methods rely on hand-engineered rewards or action-labeled demonstrations, neither of whic…
- arXiv cs.AIagentsPAIR: Prefix-Aware Internal Reward Model for Multi-Turn Agent Optimization
arXiv:2605.17877v3 Announce Type: replace Abstract: A significant hurdle for current LLMs is the execution of complex, multi-stage tasks. Group Relative Policy Optimization (GRPO) has been emerging as a leading choice, b…
- arXiv cs.AIagentsPROJECTMEM: A Local-First, Event-Sourced Memory and Judgment Layer for AI Coding Agents
arXiv:2606.12329v2 Announce Type: replace Abstract: AI coding agents often lose project-specific rationale across sessions and may repeat approaches that previously failed. The bottleneck is often not model capability bu…
- arXiv cs.AIagentsSETA: Scaling Environments for Terminal Agents
arXiv:2607.10891v2 Announce Type: replace Abstract: Large language models (LLMs) are rapidly shifting toward agents that solve tasks through diverse interfaces, including web and graphical user interfaces (GUIs). Among t…
- arXiv cs.AIagentsORACLE: Agentic AI Orchestrator Routing Via Adaptive Verifier Calibration Feedback
arXiv:2607.22465v4 Announce Type: replace Abstract: Modern enterprise agent deployments consist of a heterogeneous pool of large language models (LLMs) having diverse capabilities and cost. Existing model routing strateg…
- arXiv cs.AIagentsEnvACE: Internalizing Environment Dynamics via World Rehearsal for Agentic Reinforcement Learning
arXiv:2608.06197v2 Announce Type: replace Abstract: Training large language model agents for long-horizon tool use typically relies on interactions with real or synthesized executable environments, whose construction and…
- arXiv cs.AIagentsSAPO: Single-Rollout Autoregressive Policy Optimization for Agentic Reinforcement Learning
arXiv:2608.19842v4 Announce Type: replace Abstract: Agentic reinforcement learning (RL) has emerged as an important post-training approach for enhancing the capabilities of Large Language Models (LLMs). However, existing…
- arXiv cs.AIagentsCONTRAMEM: Learning Self-Evolving Procedural Memory from Contrasting Multi-Model Trajectories
arXiv:2608.22533v2 Announce Type: replace Abstract: Autonomous computer-use agents are increasingly applied to long-horizon tasks requiring coordinated application calls, persistent state tracking, and verifier-sensitive…
- arXiv cs.AIagentsAtlasNav: Mitigating Evidence Blindness with Persistent Corpus Navigation
arXiv:2608.24764v2 Announce Type: replace Abstract: As language-model agents become more capable of iterative search, corpus access is shifting from retrieval toward interaction. Agents can explore the corpus, inspect do…
- arXiv cs.AIagentsTowards a Belief-Based World Model for LLM Agents
arXiv:2609.00455v2 Announce Type: replace Abstract: Large language models (LLMs) are being used as policies for autonomous decision-making and planning in many domains. Despite their strong reasoning capabilities, LLMs s…
- arXiv cs.AIagentsResidualAuth: What Authorization State Must Language Agents Preserve under Revocable Delegation?
arXiv:2609.08062v2 Announce Type: replace Abstract: With revocable delegation, two histories can yield identical current permissions yet require opposite decisions for the same query after the same revocation. We introdu…
- arXiv cs.AIagentsKill-Chain Canaries: Stage-Level Tracking of Prompt Injection Across Attack Surfaces and Five Production LLMs
arXiv:2603.28013v4 Announce Type: replace-cross Abstract: Multi-agent LLM systems now read documents, web pages and tool results on behalf of users, yet their resistance to prompt injection is usually reported as one num…
- arXiv cs.AIagentsFederation over Text: Insight Sharing for Multi-Agent Reasoning
arXiv:2604.16778v3 Announce Type: replace-cross Abstract: Modern agents specialize in varying domains while there is no clear approach combining different domain skills. We propose a federated learning-like framework, Fe…
- arXiv cs.AIagentsBimanual Robot Manipulation via Multi-Agent In-Context Learning
arXiv:2604.20348v3 Announce Type: replace-cross Abstract: Large Language Models (LLMs) have emerged as powerful reasoning engines for embodied control. In particular, In-Context Learning (ICL) enables off-the-shelf, text…
- arXiv cs.AIagentsWASIL: In-the-Wild Arabic Spoken Interactions with LLMs
arXiv:2605.16364v3 Announce Type: replace-cross Abstract: Large Language Models (LLMs) voice assistants are commonly built as cascaded Automatic Speech recognition (ASR) to LLM systems, where recognition errors can disto…
- arXiv cs.AIagentsCHI-Bench: Can AI Agents Automate End-to-End, Long-Horizon, Policy-Rich Healthcare Workflows?
arXiv:2605.16679v4 Announce Type: replace-cross Abstract: End-to-end automation of realistic healthcare operations stresses three capabilities underrepresented in current benchmarks: policy density, decisions must be gro…
- arXiv cs.AIagentsDexHoldem: An Agentic Robotics Benchmark for Dexterous Manipulation in Texas Hold'em
arXiv:2605.18727v3 Announce Type: replace-cross Abstract: Evaluating embodied systems with real dexterous hardware requires more than isolated motor-skill tests: an agent must perceive a changing scene (e.g. a tabletop),…
- arXiv cs.AIagentsAdvertising to Agents: Measuring Description-Framing Bias in LLM Tool Selection
arXiv:2605.23916v3 Announce Type: replace-cross Abstract: AI agents often pick tools from registries, where each provider writes its tool's description; we ask whether sales language there changes which tool an agent pic…
- arXiv cs.AIagentsGrepSeek: Training Search Agents for Direct Corpus Interaction
arXiv:2605.29307v2 Announce Type: replace-cross Abstract: Large Language Model (LLM) search agents have shown strong promise on knowledge-intensive tasks through iterative reasoning and retrieval. Most existing systems r…
- arXiv cs.AIagentsThe Surface You Test Is Not the Surface That Breaks
arXiv:2605.30454v2 Announce Type: replace-cross Abstract: Prompt-injection benchmarks for LLM agents typically test attacks through a single injection surface and report the resulting attack success rate as a property of…
- arXiv cs.AIagentsUncertainty Quantification for Computer-Use Agents: A Benchmark across Vision-Language Models and GUI Grounding Datasets
arXiv:2606.25760v2 Announce Type: replace-cross Abstract: Computer-use agents turn vision-language model (VLM) predictions into executable GUI clicks, so reliable uncertainty estimates are essential for rejection, calibr…
- arXiv cs.AIagentsHarness Continual Learning: Continual Adaptation Beyond Model Parameters
arXiv:2608.19013v2 Announce Type: replace-cross Abstract: Continual learning has largely been model-centric, treating model parameters as the state that changes with sequential experience. Modern agents can also adapt th…