The Archive

The specification lets developer, compliance and security teams define their own policies for agents to follow in portable policy files.

Microsoft just announced "Project Solara," a new OS designed for gadgets that run AI agents, at Build 2026. The company is calling it "a new platform built from the ground up to power agent-driven experiences." It's built on Android, not Windows. Microsoft demonstrated two concept Project Solara devices at Build today: Desk concept and badge concept. The desk concept is an Amazon Echo Show-like device that unlocks with facial recognition and provides access to AI agents. The badge concept is a wearable, the type of badge you'd typically use to access a work building. It has a camera and a fin...

AI agents can increasingly handle the mechanical tasks of financial analysis: retrieving documents, calculating formulas, updating spreadsheets. The harder, more valuable challenge is reasoning through the open-ended questions that define expert Analyst work. Existing benchmarks do not capture this class of problem, and those that attempt to evaluate open-ended reasoning rely on model-judged outputs that introduce noise and circularity. We present Hedge-Bench 1.0: a benchmark of 102 actual, on-the-job tasks grounded in the explicit reasoning traces of professional hedge fund analysts working ...

Large language model (LLM) agents are evolving from request-response assistants into long-running software actors: they maintain state across model calls, fork subtasks, wait for external events, request human authority, generate tools, and perform side effects that must be resumed and audited. This paper presents Agent libOS, a library-OS-inspired runtime substrate for LLM agents. Agent libOS runs above a conventional host operating system; it does not implement hardware drivers, kernel-mode isolation, or a POSIX-compatible operating system. Instead, it treats an agent as an AgentProcess: a ...

Agent benchmarks should reflect what users actually ask deployed agents to do, yet existing benchmarks often miss key realism properties of real developer-agent sessions. We introduce RealClawBench, a live benchmark framework built from real OpenClaw sessions to capture the distribution, diversity, and real-world difficulty of deployed agent use. Real user requests are challenging to benchmark because they often depend on local execution environments, involve implicit or underspecified intent, and require nontrivial verification. RealClawBench addresses these challenges with two core mechanis...

GitHub pioneered the modern AI coding era with Copilot, and the resulting explosion in agentic coding has led to notable strains on the most popular developer platform in the world. Here's the plan.

Multi-Document Summarization (MDS) plays a critical role in distilling essential information from collections of textual data. Existing approaches often struggle to capture complex inter-document relationships, rely heavily on large amounts of labeled data for supervised training, or exhibit limited generalization across domains and languages. To address these limitations, we present a training-free mixture-of-agents framework for MDS that leverages the complementary strengths of large language models (LLMs) and knowledge graphs. Our approach decomposes summarization into specialized agent ta...

Recent progress in Large Language Model (LLM) agents has enabled promising advances in automated data science. However, existing approaches remain fundamentally limited by their static action sets and lack of principled long-horizon context management, hindering their ability to accumulate reusable experience across tasks and operate reliably in multi-stage, iterative data science pipelines. To address these challenges, we introduce EvoDS, a self-evolving autonomous data science agent that learns to expand its skills and adaptively managing long-term context through agentic reinforcement lear...

Financial-research answers are decision-relevant only when another analyst can audit how they were produced: which source was chosen, which period and accounting definition were used, which assumptions were made, and how the calculation was performed. Existing finance benchmarks largely evaluate isolated subskills or final answers, leaving the auditable derivation itself under-measured. We introduce BigFinanceBench, a 928-item expert-authored benchmark of open-ended financial-research tasks in which each item pairs a ground-truth reference answer with a point-weighted rubric that decomposes t...

AI agents are a powerful tool for synthesizing data to accelerate research, summarize information, and help teams make decisions faster. But combining internal... AI agents are a powerful tool for synthesizing data to accelerate research, summarize information, and help teams make decisions faster. But combining internal data with public sources poses security challenges. This post shares an open source example using Hermes Agent with NVIDIA NemoClaw for product research across Outlook, Slack, and GitHub. NVIDIA OpenShell enforces a security-approved… Source

AI-assisted coding agents are bottlenecked by input-token cost. Two pathologies of raw human input drive much of this overhead: tokenization inefficiency for non-English text and structural entropy in conversational prompts. Existing approaches act reactively by compressing already-bloated contexts or intervening after failures occur. We introduce a pre-flight, edge-side prompt-rewriting middleware that operates between the developer and the cloud agent. A local Llama 3.2 (3B) model performs cross-lingual translation into English, structural rewriting into a compact task-oriented format, and ...

Embodied agents that navigate cities rely on world models that predict how their surroundings will change as they move. But for navigation, what matters is not what the buildings look like; it is where the agent can go. Most world models nonetheless predict appearance, learning how a scene looks rather than the space an agent can move through. Those that do target geometry, such as bird's-eye-view occupancy grids, flatten the three-dimensional environment onto a ground plane, discarding the above-ground and multi-level structure that shapes real navigation. What is missing is a predictive tar...

Self-improving language agents are typically evaluated in isolation: an agent attempts a task, receives feedback, and iteratively refines its own behavior. Yet agents increasingly operate alongside peers whose strategies and outcomes are publicly visible. This raises an under-studied question: when does shared experience produce improvements that self-improvement alone cannot achieve? We introduce SAGE (Social Agent Group Evolution),an evaluation framework that compares two compute-matched conditions: SocialEvo, where agents from five distinct model families co-evolve with access to all peers...

To improve the real-world applicability of reinforcement learning (RL), the field of adversarially robust RL studies how to train agents under adversarial environment perturbations. In this setting, a protagonist agent optimizes a policy under environmental perturbations from an adversary, resulting in a zero-sum Markov game. When adversarially robust RL is combined with model-based RL, the adversary can target a learned transition model instead of the training environment. Extending this idea, this work introduces post-hoc robustification of deep RL agents at inference time. By using the lea...

As AI systems evolve from passive models into autonomous active agents capable of initiating actions, collaborating, and delegating tasks, the traditional boundaries of software systems blur. Traditional authorization and delegation frameworks, built around fixed principals, explicit requests, and static scopes, are insufficient to govern agentic systems. Agentic AI demands richer authorization semantics: agents must inherit and delegate permissions, act under time-limited authority, and coordinate through shared protocols. Existing Identity and Access Management (IAM) systems fail to fully c...

Conversational AI agents require memory systems that are both scalable and semantically coherent across long interaction horizons. Existing approaches rely predominantly on large language model (LLM)-based summarisation at write time, which introduces non-determinism, escalating token costs, and opacity in pruning decisions. We present the Deterministic Memory Framework (DMF), a CPU-first approach that replaces generative memory compression with a fully deterministic pipeline grounded in classical NLP analysis, vector geometry, and mathematical scoring. DMF assigns each conversational interac...

Stronger code agents are commonly assumed to be superior teachers for post-training, yet this assumption remains poorly disentangled from task difficulty, harness design, and student capacity. We investigate this pedagogical link using Terminal-Lego, a scalable pipeline that transforms multi-domain real-world issues into environment-verified agentic tasks. Surprisingly, standalone performance does not dictate teaching efficacy: while Claude Opus 4.6 achieves higher scores on Terminal-Bench 2.0, students fine-tuned on trajectories from DeepSeek-V3.2, a lower-scoring agent, exhibit significantl...

Vulnerability disclosure volumes now far exceed organizational assessment capacity, yet three adjacent research communities (proof-of-concept generation, vulnerability prioritization, and detection rule engineering) operate largely in isolation. Existing automated exploit generation systems report binary pass/fail outcomes, discarding partial progress and producing no signal for the other two communities. This paper presents FORGE, a multi-agent system that bridges these three silos through graduated exploitation depth. Five specialized agents (Intel, Generator, Planner, Exploit, and Detector...

As AI agents move from the digital world to the physical environment, they can readily use NVIDIA Jetson to accelerate real-world deployment with optimized... As AI agents move from the digital world to the physical environment, they can readily use NVIDIA Jetson to accelerate real-world deployment with optimized memory and performance. NVIDIA JetPack 7.2 directly supports one-command deployment of NVIDIA NemoClaw, an open source stack that adds privacy and security controls to OpenClaw. It introduces NVIDIA agent skills for Jetson—Jetson device… Source

The rise of autonomous, long-running AI agents has introduced a new class of compute demand, namely tasks that maintain large context windows, spawn concurrent... The rise of autonomous, long-running AI agents has introduced a new class of compute demand, namely tasks that maintain large context windows, spawn concurrent subagents, and iterate continuously without cloud dependency. Security and privacy concerns are also accelerating the shift toward local agents. Developers, by running autonomous agents on hardware they own with NVIDIA NemoClaw… Source

If Nvidia has cracked a way to bring AI agents easily, safely and usefully to the masses, it could — and should — be big.

Clinical practice is not the selection of an answer from enumerated options: a physician gathers heterogeneous information incrementally and commits to sequential, irreversible decisions under uncertainty. Static benchmarks cannot probe and existing interactive medical benchmarks each compromise on at least one of them. We present ClinEnv, an interactive benchmark that evaluates LLMs as attending physicians over real inpatient admissions under a paradigm we term Longitudinal Inpatient Simulation. Each case is automatically constructed into an ordered sequence of decision stages; at every stag...

We introduce HERO'S JOURNEY, a benchmark for rule induction in goal-directed episodic tasks, where agents must infer hidden rules from demonstrations and act on them through multi-step execution. HERO'S JOURNEY covers eight tasks across attribute and procedural induction families, each with four structural rule forms, controllable lexical grounding, and identifiability conditions. Evaluating state-of-the-art LLMs, we find that models show evidence of rule induction, but the ability is limited and uneven across tasks. Meanwhile, process execution adds an execution bottleneck for models, wherea...

Agent skills occupy a privileged position in the agent workflow, as agents are expected to implicitly follow and execute them, rendering third-party skills a vulnerable attack surface. Existing studies have revealed unsafe agent behaviors induced by skill-based attacks, but they primarily evaluate poisoned skills within a single task execution and enumerate harms through ad-hoc risk lists. To bridge these gaps, we introduce SkillHarm, a benchmark of skill-based attacks across the skill-use lifecycle, paired with a systematic taxonomy of skill-relevant risks. SkillHarm evaluates two attack sce...

Text files such as skill files, memory files, and behavioral configuration files play a central role in defining how modern agents act. Through edits by humans or the agents themselves, these files may evolve over time, directly steering the agent's behavior in future interactions. We present a methodology and framework for measuring agent $traits$ by defining traits as directions in the embedding space of a text embedding model. We train a linear model on labeled "before" versus "after" skill file diffs to learn a trait vector, then score arbitrary skill edits by projecting their embedding d...

Large language models now power robo-advisors and trading agents, yet whether they carry built-in biases toward specific assets is largely untested. We ask three questions: do LLMs systematically prefer certain financial instruments; can an internal representation with causal leverage over those preferences be identified; and does that representation affect downstream financial decisions? We develop a three-level audit protocol and apply it to Bitcoin. First, a behavioral audit of eight frontier LLMs shows that Bitcoin's ranking among money-like instruments is frame-dependent: models place it...

Time series forecasting has advanced rapidly, especially with the emergence of foundation models that show strong zero-shot performance on numerical extrapolation. However, in real-world forecasting settings, a statistically plausible baseline is rarely the final forecast used in practice. Before a forecast becomes decision-ready, it often needs to be revised using weakly structured business context such as holiday effects, campaign plans, external events, historical analogs, and expert feedback. This practical stage remains underexplored in the forecasting literature. In this paper, we formu...

Tool-augmented language agents speculatively issue likely future tool calls to hide latency, but those calls leak inferred user intent to external services before the agent commits to the branch. Every external observer that received the call retains the disclosure after the agent abandons the branch. Timing is the issue, not authorization: no commit-time cleanup, read-only restriction, or access-control allow-list unsends what an observer already holds. We call these invocations ghost tool calls and propose Speculative Tool Privacy Contracts, a runtime abstraction that treats observation bef...

The Model Context Protocol (MCP) has emerged as a transformative standard for connecting large language models (LLMs) with external data sources and tools, and has been rapidly adopted across personal applications and development platforms. However, existing benchmarks predominantly focus on generic information-seeking tools and fail to capture the practical challenges posed by personal social applications, where tools interact with individual accounts or local databases. To bridge this critical gap, we introduce MCP-Persona, the first benchmark specifically designed for evaluating agent perf...