今日必看

Approval Laundering: Systematizing Approval--Execution Binding Failures in AI Coding-Agent Harnesses

arXiv:2609.38983v1 Announce Type: cross Abstract: Modern AI coding-agent harnesses (Claude Code, Codex CLI, Cursor) rest their security boundary on a largely unexamined assumption: that the action A a human approves is the same action A' the harness executes, where A is fixed by

Argus: Academic Integrity in the Era of Generative AI

arXiv:2609.36073v1 Announce Type: cross Abstract: The rapid proliferation of large language models (LLMs) in the context of education has introduced significant challenges in enforcement of academic integrity, especially in programming courses. We present Argus, an automated dete

Cheap to Hypothesize, Costly to Verify: The Defense Surface of Agentic Vulnerability Discovery

arXiv:2609.35909v2 Announce Type: replace-cross Abstract: Autonomous LLM agents turn vulnerability discovery into a repository-scale search: they generate many vulnerability hypotheses but can verify only a subset under a finite budget. We show that autonomous vulnerability disco

Composing Task-specific Agent Harnesses at Test Time with Reusable Primitives

arXiv:2609.38912v1 Announce Type: new Abstract: Agent harnesses govern how large language models (LLMs) gather context, invoke tools, verify results, preserve state, and terminate, largely affecting agent performance. However, the value of each harness mechanism can differ across

Does This Action Still Explain the Task? Reverse Scoring for Diffusion Language Model Agents

arXiv:2609.38536v1 Announce Type: cross Abstract: Diffusion-based large language models (dLLMs) promise to break the sequential latency bottleneck of autoregressive agents through parallel decoding, but recent evaluations show this efficiency does not transfer to embodied agentic

E2E-SWE: Benchmarking LLMs on Building Working Codebases from Scratch

arXiv:2609.38335v1 Announce Type: cross Abstract: Coding agents powered by large language models (LLMs) are evolving from making localized code changes to developing complete software repositories. However, evaluating repository-scale generation remains challenging: tasks must de

GraphCert: Bootstrap Agentic Graph Reasoning with Certified Evidence Rubrics

arXiv:2609.38798v1 Announce Type: new Abstract: Graph agents extend large language models (LLMs) with the ability to actively explore and reason over knowledge graphs through multi-step interactions with graph tools. However, training capable graph agents typically requires large

Harness as a Language: A Minimalist Agent Framework With Maximal Expressivity

arXiv:2609.26891v2 Announce Type: replace Abstract: Modern language-model agents are built around the agent loop: the LLM is placed in an environment exposing a set of tools, and the LLM has full control over the workflow by alternating between tool calls and observing their outp

Inference Auctions

arXiv:2609.40070v1 Announce Type: cross Abstract: When inference demand exceeds available compute capacity, model providers must decide which requests should be served first. Users have different tolerances for delay from an LLM API, but current priority pricing schemes compress

Instability Floors: Separating Bias from Noise in Fairness Audits of Clinical LLM Agents with FairMedAgent

arXiv:2609.03221v3 Announce Type: cross Abstract: Counterfactual fairness audits of clinical language-model agents report a flip rate: how often an action changes when only the patient's demographic descriptor changes. Part of that rate is not demographic. A stochastic agent also

Janus: Evidence-Before-Effect Sagas and Offline-Verifiable Provenance for Agentic LLMs

arXiv:2609.38266v1 Announce Type: cross Abstract: Agentic large language models (LLMs) now move money through tools, yet the record of what they did is usually a trace their own process emits beside the effect. Janus puts the record on the effect path. A step's proposal, the verd

Learning to Route in Visual Space via Multi-Step Embedding Retrieval

arXiv:2609.38743v1 Announce Type: new Abstract: LLM agents rely on retrieval tools to access external knowledge, yet visual agentic search remains severely bottlenecked by standard single-step retrievers. In current pipelines, the agent must issue text queries for every intermedi

LongEmo: Towards Emotion Understanding and Reasoning in Long Videos

arXiv:2609.40079v1 Announce Type: cross Abstract: While recent Multimodal Large Language Models (MLLMs) have shown promise in affective computing, their reasoning capabilities are largely confined to short video clips with limited interactions. However, real-world emotions are no

MiniRep: Robust Reputation-Based Aggregation for Multi-Agent Debate

arXiv:2609.39297v1 Announce Type: new Abstract: Autonomous agents powered by large language models (LLMs) are rapidly evolving into an open agentic ecosystem. To support trustworthy collaboration, industry initiatives increasingly assess agent reputation from past behavior and pr

ORACLE: Agentic AI Orchestrator Routing Via Adaptive Verifier Calibration Feedback

arXiv:2607.22465v3 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 strategies optimize the quality-cost trade-off, while providing req

Personalized State-Transition-Aware Memory for Clinical Agents

arXiv:2609.38490v1 Announce Type: cross Abstract: Large language model (LLM) agents that reason over clinical records must track changes in a patient's state while preserving the history needed to understand them. Simply accumulating memories leaves it unclear which information s

RankEvolve: A Reliable Multi-Agent Auto-Research Harness for Evolving Ranking Models

arXiv:2609.39551v1 Announce Type: new Abstract: Auto-research agents, LLM systems that propose, implement, train, and evaluate model changes across iterations, promise to automate applied ML's experimental loop. Over long horizons, execution accuracy is a binding constraint: a ch

Ranking-Aware Prompt Optimization for Multimodal Clinical Diagnosis

arXiv:2609.40361v1 Announce Type: cross Abstract: Multimodal large language models (MLLMs) are rapidly advancing clinical diagnosis, yet their adaptation pipelines remain anchored to accuracy-based objectives. Clinical data are heavily class-imbalanced: a constant-majority predic

SAPO: Single-Rollout Autoregressive Policy Optimization for Agentic Reinforcement Learning

arXiv:2608.19842v3 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 methods face a trade-off between policy performance and res

SEABench: Benchmarking Endogenous Misalignment In Self-Evolving Agents

arXiv:2609.35596v2 Announce Type: replace-cross Abstract: Self-evolving LLM agents have gained prominence for their ability to improve after deployment by modifying their harness, including their controller instructions, memory management protocols, and reusable tools and skills,

Schema: Discovering Unknown Environments via Agentic Program Induction

arXiv:2609.39140v1 Announce Type: new Abstract: Learning to complete tasks in unfamiliar environments with unknown rules remains a key challenge for LLM agents. Current LLM agents often record their discoveries in prose, which may not provide a compact, explicit account of how th

FLUX 3 Image 上线 OpenRouter，支持原生 4K 生成与多参考编辑

Black Forest Labs 的 FLUX 3 Image 现已上线 OpenRouter，是支持文生图与多参考编辑的旗舰图像模型，原生可渲染至 4K。原文提到可精确多轮编辑不动其他像素、用 bounding box 布局、最多用 10 个参考图合成，商业权重已开放，开放权重版将在未来数周发布。 🔗 阅读原文 via AIHOT · https://aihot.news/items/x1x0d5mqj4y384t9fcq4415kj

AI agents can now buy real services

Where We Expect AMD EPYC 9006 CPUs in the Era of Agentic AI – ServeTheHome