相关论文: Do Coding Agents Understand Least-Privilege Author…

We address the problem of learning to assign prediction tasks to one agent from a set of available human or AI agents. In particular, we focus on the sequential learning of agent expertise and assignment policies where each agent is…

This article, a lightly adapted version of Perplexity's response to NIST/CAISI Request for Information 2025-0035, details our observations and recommendations concerning the security of frontier AI agents. These insights are informed by…

Since autonomous coding agents generate complex behaviors at high-volume, we may want to use other LLMs to monitor actions to reduce the risk from dangerous misaligned behavior. To better understand the limitations of such monitors against…

Agent systems often decompose a task across multiple roles, but these roles are typically specified by prompts rather than enforced by access controls. Without enforcement, a team pass rate can mask whether agents actually coordinated or…

Affordance grounding requires identifying where and how an agent should interact in open-world scenes, where actionable regions are often small, occluded, reflective, and visually ambiguous. Recent systems therefore combine multiple skills…

Colluding language-model agents can hide coordination in messages that remain policy-compliant at the surface level. We present CLBC, a protocol where generation and admission are separated: a message is admitted to transcript state only if…

Algorithm designers increasingly optimize not only for accuracy, but also for the fairness of the algorithm across pre-defined groups. We study the tradeoff between fairness and accuracy for any given set of inputs to the algorithm. We…

Recent work has shown that many problems of satisfiability and resiliency in workflows may be viewed as special cases of the authorization policy existence problem (APEP), which returns an authorization policy if one exists and 'No'…

Access control is fundamental to computer security, and has thus been the subject of extensive formal study. In particular, *relative expressiveness analysis* techniques have used formal mappings called *simulations* to explore whether one…

Robust access control is a cornerstone of secure software, systems, and networks. An access control mechanism is as effective as the policy it enforces. However, authoring effective policies that satisfy desired properties such as the…

Authorizing Large Language Model driven agents to dynamically invoke tools and access protected resources introduces significant risks, since current methods for delegating authorization grant overly broad permissions and give access to…

Real-world requests to AI agents are fundamentally underspecified. Natural human communication relies on shared context and unstated constraints that speakers expect listeners to infer. Current agentic benchmarks test explicit…

Users often interpret and select agent skills through their SKILL markdown specifications. To protect users, existing audits mainly focus on malicious or unsafe skills. We study the complementary question of whether specifications help…

Recent work has demonstrated the plausibility of frontier AI models scheming -- knowingly and covertly pursuing an objective misaligned with its developer's intentions. Such behavior could be very hard to detect, and if present in future…

Large Language Model (LLM) agents demonstrate strong performance in autonomous code generation under loose specifications. However, production-grade software requires strict adherence to structural constraints, such as architectural…

Model stealing, i.e., unauthorized access and exfiltration of deep learning models, has become one of the major threats. Proprietary models may be protected by access controls and encryption. However, in reality, these measures can be…

When should an autonomous agent commit resources to a task? We introduce the Agent Capability Problem (ACP), a framework for predicting whether an agent can solve a problem under resource constraints. Rather than relying on empirical…

As Model Context Protocol adoption grows, securing tool invocations via meaningful user consent has become a critical challenge, as existing methods, broad always allow toggles or opaque LLM-based decisions, fail to account for dangerous…

LLM-powered agents often use prompt compression to reduce inference costs, but this introduces a new security risk. Compression modules, which are optimized for efficiency rather than safety, can be manipulated by adversarial inputs,…

AI agents today have passwords but no permission slips. They execute tool calls (fund transfers, database queries, shell commands, sub-agent delegation) with no standard mechanism to enforce authorization before the action executes. Current…