AI news archive.
4,965 articles filed under Agents, newest first, from the outlets listed on the sources page.
- arXiv cs.AIagentsAPI Secrets Should Never Become Tokens in the LLM's Vocabulary: A Threat Analysis of API Credential Handling in LLM Agent Systems and an Empirical Evaluation of a Vault-Mediated Execution Boundary
arXiv:2609.33371v1 Announce Type: cross Abstract: Tool-using large language model (LLM) agents turn credential hygiene from a storage problem into an execution-security problem. A key pasted into a prompt, or embedded in…
- arXiv cs.AIagentsRecursive Harness Distillation across Agents for Robot Manipulation
arXiv:2609.33378v1 Announce Type: cross Abstract: A central goal in robotics is to enable manipulation across changing tasks and environments. Vision-language-action (VLA) models provide broad manipulation capabilities b…
- arXiv cs.AIagentsNLPG: Natural-Language Policy Gradients for Self-Evolving Language Agents
arXiv:2609.33379v1 Announce Type: cross Abstract: Large language model agents increasingly rely on compound programs for retrieval, tool use, reasoning, and verification, yet their failures often arise from local procedu…
- arXiv cs.AIagentsBeyond Timestamps: Decision-Aligned On-Policy Distillation for Long-Horizon Agents
arXiv:2609.33391v1 Announce Type: cross Abstract: Reinforcement learning with verifiable rewards (RLVR) often relies on sparse outcome rewards, providing coarse supervision for long-horizon agents. On-policy self-distill…
- arXiv cs.AIagentsEvaluating System One Models for Agent Security Decisions: Reliability, Calibration, and Selective Automation
arXiv:2609.33401v3 Announce Type: replace-cross Abstract: Agentic software connects language models to tools that modify files and interact with external services. Developers use model-based security checks to screen ext…
- arXiv cs.AIagentsDataMagic: Authoring Data Videos through Declarative Multi-Agent Orchestration
arXiv:2609.33403v1 Announce Type: cross Abstract: Data videos communicate data insights through dynamic charts, voice narration, and synchronized animations, and have become a widely adopted form of data storytelling. Ho…
- arXiv cs.AIagentsGraph-Guided Repository Environment Construction
arXiv:2609.33429v1 Announce Type: cross Abstract: Coding agents now increasingly rely on execution to validate their solutions, making the construction of reliable execution environments a critical enabling capability. H…
- arXiv cs.AIagentsHESP: Separating What to Probe from When to Stop in Local LLM Alert-Triage Agents
arXiv:2609.33446v1 Announce Type: cross Abstract: Security operations centers receive far more alerts than analysts can investigate, and organizations that cannot send their telemetry to hosted models must automate triag…
- arXiv cs.AIagentsMA-JEPA: Joint-Embedding World Models for Multi-Agent Reinforcement Learning
arXiv:2609.33563v2 Announce Type: replace-cross Abstract: World models improve sample efficiency by training policies on imagined trajectories, but their usefulness depends on learning representations that capture the in…
- arXiv cs.AIagentsStoryEngine: A State-Grounded Agentic Framework for Video Storytelling
arXiv:2609.33627v1 Announce Type: cross Abstract: Despite recent progress in agentic multi-shot video generation, producing coherent and consistent long-form stories remains challenging. Existing agentic pipelines typica…
- arXiv cs.AIagentsSecProbe: Adaptive Evaluation of Coding Agents on Cybersecurity Vulnerabilities
arXiv:2609.33763v1 Announce Type: cross Abstract: Assessing cybersecurity vulnerability awareness in coding agents requires evaluations that reveal capability gaps and remain informative as models evolve. Static benchmar…
- arXiv cs.AIagentsDEALS: Decentralized Expertise-Aware Load Serving for Multi-Agent LLM Systems
arXiv:2609.33768v1 Announce Type: cross Abstract: Multi-agent systems (MAS) have recently emerged as an effective approach for coordinating large language model (LLM)-based agents to solve complex tasks through structure…
- arXiv cs.AIagentsCodeActionBench: Evaluating Agentic Code-as-Policy for Embodied Manipulation
arXiv:2609.33807v1 Announce Type: cross Abstract: How well can general-purpose multimodal models turn visual understanding and reasoning into embodied manipulation via executable code? We introduce CodeActionBench, a ben…
- arXiv cs.AIagentsCounterfactual Rollout Replay: Forkable Environments as Free Process Rewards for Software Engineering Agents
arXiv:2609.33875v3 Announce Type: replace-cross Abstract: Outcome-only reinforcement learning gives software engineering (SWE) agents a terminal success signal but little direct guidance about intermediate decisions. We…
- arXiv cs.AIagentsMaat: Independent Deterministic Contract-Based Governance for Multi-Agent LLM Workflows
arXiv:2609.34017v1 Announce Type: cross Abstract: Large-language-model multi-agent systems (LLM-MAS) introduce a characteristic reliability problem: an error produced by one agent can be accepted as context by downstream…
- arXiv cs.AIagentsUOPD: Uncertainty-Aware Intervention for On-Policy Distillation of Multi-Turn Agents
arXiv:2609.34036v1 Announce Type: cross Abstract: On-policy distillation (OPD) trains a student on its own rollouts using dense supervision from a teacher. In multi-turn environments, a mistake at a critical decision ste…
- arXiv cs.AIagentsLearning Perturbation Robust Policies for LLM Agents with Stable Optimization
arXiv:2609.34064v1 Announce Type: cross Abstract: Reinforcement learning (RL) has become an effective post-training paradigm for long-horizon large language model (LLM) agents. However, we find that the resulting policie…
- arXiv cs.AIagentsAD-E2E-JEPA: A Joint-Embedding Predictive Architecture For End-to-End Autonomous Driving
arXiv:2609.34085v1 Announce Type: cross Abstract: Autonomous driving requires \textit{world models} that can understand the physical world, reason and plan, and operate safely. In this paper, we first systematically eval…
- arXiv cs.AIagentsJET: Judge-Guided Evolution at Test Time for Agent Programs
arXiv:2609.34126v1 Announce Type: cross Abstract: An agent's executable program governs how it uses tools, processes observations, and responds to failures. Evolving this program at test time can help adaptation, but dec…
- arXiv cs.AIagentsCertified Multi-Source Integrity for Structured Agent Actions
arXiv:2609.34245v1 Announce Type: cross Abstract: LLM agents increasingly take privileged, often irreversible structured actions, such as paying an invoice. They assemble each action from action-critical fields in docume…
- arXiv cs.AIagentsSAGE: Symbolic Action-Gating and Editing for LLM Task Planners
arXiv:2609.34268v1 Announce Type: cross Abstract: Large language models (LLMs) are now the default cognitive core of embodied household agents, yet the plans they emit are rarely checked against a grounded model of the e…
- arXiv cs.AIagentsDr.Credit: Rubric-Grounded Process Credit Assignment for Deep Research Agents
arXiv:2609.34296v1 Announce Type: cross Abstract: Rubric-based tasks are increasingly addressed through reinforcement learning (RL), with rubric scores used as training rewards. However, these rewards typically supervise…
- arXiv cs.AIagentsSAIL: Spatial Audio Intelligence with Large Language Models via Disentangled Acoustic-Spatial Encoding and Dual-Stream Q-Former
arXiv:2609.34347v1 Announce Type: cross Abstract: Spatial audio large language models (LLMs) enable embodied agents, wearable assistants, and immersive systems to recognize sound events, localize sources, and reason abou…
- arXiv cs.AIagentsBefore Agents Act: Assurance-Aware Semantic Scheduling for Evidence Acquisition in Distributed Systems
arXiv:2609.34376v1 Announce Type: cross Abstract: Tool-using agents can initiate consequential infrastructure changes, yet evidence required for admission may expire while other checks run or depend on a shared fault dom…
- arXiv cs.AIagentsJust-In-Time Agent Memory with Runtime Agentic Research
arXiv:2609.34385v1 Announce Type: cross Abstract: Memory is critical for AI agents. Many existing agent-memory systems follow an Ahead-of-Time (AOT) design, constructing memory before a specific request arrives. While th…
- arXiv cs.AIagentsCoding Agent Memory Post-training: Unlocking the Memory Potential of Pre-trained File Operations for Long-Horizon Tasks via Reinforcement Learning
arXiv:2609.34422v2 Announce Type: replace-cross Abstract: Language-model agents increasingly tackle long-horizon tasks whose interaction histories exceed the model's active context. Recent work has begun to use reinforce…
- arXiv cs.AIagentsAgentHop: A Diagnostic Benchmark for Agentic Multi-Hop Scientific Question Answering
arXiv:2609.34428v1 Announce Type: cross Abstract: Agentic tasks require a large language model to interact with the world, navigating information and gathering evidence across multiple steps with restricted resources. Du…
- arXiv cs.AIagentsRemember by Asking: Retrieval-Induced Memory Evolution for LLM Agents
arXiv:2609.34438v1 Announce Type: cross Abstract: Long-term memory is essential for language agents to maintain coherent and effective behavior over extended, multi-session interactions. Existing memory systems mainly us…
- arXiv cs.AIagentsCoDeL: Co-Evolutionary Defense against Indirect Prompt Injection in LLM-based Agents
arXiv:2609.34463v1 Announce Type: cross Abstract: Large language model (LLM)-based agents increasingly rely on external tools and content, exposing them to indirect prompt injection (IPI). This threat has motivated a wid…
- arXiv cs.AIagentsSEAD: A State-Based Perspective on Attack and Defense in Tool-Using Agents
arXiv:2609.34518v1 Announce Type: cross Abstract: Language-model agents increasingly use tools to act on external systems. Earlier actions can alter files, permissions, database records, or other state, making a later ro…
- arXiv cs.AIagentsCodoku: Renewable Program-Reasoning Challenges for Frontier Coding Agents
arXiv:2609.34661v1 Announce Type: cross Abstract: Existing program-reasoning benchmarks ask large language models to predict a program's behavior on a given input. Coding agents break two assumptions on which these bench…
- arXiv cs.AIagentsFrom Preference to Reciprocity: Decentralized Matching with Empirically Grounded LLM-agent Based Modeling
arXiv:2609.34679v1 Announce Type: cross Abstract: Bipartite matching is a fundamental problem in game theory and market design. Classical approaches such as Gale--Shapley assume complete preferences and centralized compu…
- arXiv cs.AIagentsCoDrive: Cross-Vehicle World-Consistent Video Generation with Precise Trajectory Control for Cooperative Driving
arXiv:2609.34749v1 Announce Type: cross Abstract: Real-world driving is inherently multi-agent, yet most existing driving world models generate observations from a single ego vehicle. Independently extending them to mult…
- arXiv cs.AIagentsLongPuzzleBench: Evaluating GUI Agents on Long-Horizon Visual Puzzles
arXiv:2609.34769v1 Announce Type: cross Abstract: GUI agents need long-horizon visual reasoning: they must interpret a changing interface while keeping a multi-step plan viable as earlier actions constrain later ones. Ex…
- arXiv cs.AIagentsCoSec: Benchmarking Agent Security in Communities
arXiv:2609.34790v4 Announce Type: replace-cross Abstract: LLM agents operate in persistent collaborative environments involving multiple users, communities, memories, files, and tools. Community boundaries may remain fix…
- arXiv cs.AIagentsDivOPD: Spread Wide, Look Close for Asynchronous On-Policy Distillation of Multi-turn Agents
arXiv:2609.34838v1 Announce Type: cross Abstract: On-policy distillation (OPD) trains student agents through teacher supervision on their own interactions with an environment. However, in asynchronous multi-turn training…
- arXiv cs.AIagentsAttention-based Hierarchical Variational Information Bottleneck for Robust Multi-Agent Communication under Variable Bandwidth
arXiv:2609.34860v1 Announce Type: cross Abstract: Learning-based multi-agent communication under limited bandwidth does not only require deciding what to communicate, but also structuring messages so that partial transmi…
- arXiv cs.AIagentsAudit the Scaffold, Not the Checkpoint: A Stationarity Dichotomy for Recursive Self-Improvement in Agentic Coding
arXiv:2609.34924v1 Announce Type: cross Abstract: An auditor who checks whether a system's weights are frozen is checking the wrong thing. Our stationarity dichotomy says that iterative self-modification hits strict dimi…
- arXiv cs.AIagentsPEAR: Progressive Evidence-Based AutoResearch for Industrial Search Systems
arXiv:2609.35031v1 Announce Type: cross Abstract: AutoResearch improves systems through iterative experimentation: agents propose candidate modifications, evaluate them, and use the results to guide subsequent exploratio…
- arXiv cs.AIagentsTIDE: Teacher-Student Transition via Informative Distillation and Exploration for Agentic RL
arXiv:2609.35058v1 Announce Type: cross Abstract: Effective multi-turn agents require interaction strategies that coordinate information gathering, actions, and feedback over long horizons. GRPO is a reinforcement learni…
- arXiv cs.AIagentsTimeline-Bench: Evaluating Agents on Realistic Video-Editing Tasks, from Raw Footage to Final Cut
arXiv:2609.35143v1 Announce Type: cross Abstract: AI agents increasingly carry out long-horizon professional work, but their evaluations rarely require a finished creative deliverable. To this end, we introduce Timeline-…
- arXiv cs.AIagentsResearch-Native by Construction: Minimal Nodes, Re-verifiable Workflows, and Compounding Memory for Long-Horizon Scientific Agents
arXiv:2609.35182v1 Announce Type: cross Abstract: We describe AfS (Agent for Science), a platform built for long-horizon scientific work, where a project runs for tens of hours across dozens of agent runs with a human pr…
- arXiv cs.AIagentsTeacher-Student Gaps Are Not Enough: Outcome-Guided On-Policy Distillation for Multi-Turn Autonomous Agents
arXiv:2609.35319v1 Announce Type: cross Abstract: On-policy distillation (OPD) trains a student on its own trajectories with dense teacher supervision. Recent work on OPD for multi-turn autonomous agents often treats lar…
- arXiv cs.AIagentsDo Coding Agents Reuse Existing Code or Reinvent the Wheel?
arXiv:2609.35357v1 Announce Type: cross Abstract: Coding agents are increasingly deployed for iterative development on real repositories, yet existing evaluation barely answers a basic question: \emph{do coding agents re…
- arXiv cs.AIagentsPlanarian: Managing Agent State with Statepoints
arXiv:2609.35366v1 Announce Type: cross Abstract: LLM agents solve complex tasks by iteratively changing files, invoking local tools, and interacting with remote services, which modifies state across their local environm…
- arXiv cs.AIagentsMultilinguality in Hybrid Attention LLMs
arXiv:2609.35378v1 Announce Type: cross Abstract: In response to the growing demand for long sequences in agentic and reasoning use cases, many state-of-the-art LLMs combine multiple variants of attention to mitigate the…
- arXiv cs.AIagentsMCP Error Messages Written for Developers Hurt the Most Capable Agents Most
arXiv:2609.35381v2 Announce Type: replace-cross Abstract: Many Model Context Protocol (MCP) servers wrap web APIs built for human developers, and their error messages tell the reader to run a command, edit a configuratio…
- arXiv cs.AIagents"Nothing to See Here'': Unintended Disclosure through Revision Traces of LLM Deliverables
arXiv:2609.35408v1 Announce Type: cross Abstract: Large language model (LLM) assistants increasingly help users draft content for third-party recipients. During private drafting, the user or the model may introduce an it…
- arXiv cs.AIagentsAutoRef: Harness Optimization for Agentic Multi-Reference Image Generation
arXiv:2609.35530v1 Announce Type: cross Abstract: Recent image generation models can take multiple reference images as input and combine them into a new image. However, multi-reference image generation remains challengin…
- arXiv cs.AIagentsThe Compiler May Read It, the Agent May Not: Keeping Part of a Research Code Away from a Coding Agent
arXiv:2609.35557v2 Announce Type: replace-cross Abstract: The compiler must read modules a physics-based solver cannot build without; the coding agent must not read that intellectual property. The harness, the software a…