Newest papers

AgentPolicyForge: From Executable Environments to Policy Benchmarks and Adaptive Guards

Chao Shang, Wanyu Du, Yi Zhang, Sandesh Swamy · 2026 · arXiv (Cornell University)

DOI: 10.48550/arxiv.2610.28989 · cited by 0 works (OpenAlex) · open-access PDF

Portuguese

As implementações de agentes fixos personalizados definem políticas sobre ferramentas, recursos e condições de execução. Fornecer essas políticas na mensagem do agente de tarefas não impede de forma confiável violações. Treinar um guarda requer exemplos de chamadas de ferramentas que cumprem ou violam essas políticas. Introduzimos o AgentPolicyForge, um framework que usa ambientes executáveis para construir benchmarks de políticas e treiná-las de forma iterativa. Com este framework, construímos o PolicyGym, um conjunto de dados em dez famílias de tarefas de políticas usando 698 cenários para a geração de dados de treinamento e 254 cenários separados para testes de testes avaliados por humanos. As chamadas de ferramentas são executadas e seus resultados registrados para ajudar a determinar se elas cumprem a política.

Machine translation (Meta M2M100); the English original follows.

Abstract

Customized agent deployments define policies over tools, resources, and execution conditions. Providing these policies in the task agent's prompt does not reliably prevent violations. Training a guard requires examples of tool calls that comply with or violate those policies. We introduce AgentPolicyForge, a framework that uses executable environments to construct policy benchmarks and iteratively train policy guards. With this framework, we construct PolicyGym, a dataset across ten policy families using 698 scenarios for training data generation and 254 separate held-out scenarios for human-reviewed test sets. Tool calls are executed and their outcomes recorded to help determine whether they comply with the policy. Every violating case uses a permitted tool, requiring checks of arguments, resources, and context rather than tool identity alone. The case construction process supports iterative guard training: an attacker generates requests using retrieved strategies, and the guard is fine-tuned on balanced, environment-verified examples. Training improves policy enforcement even with a fixed archive. Updating the archive further reduces in-domain attack success compared with keeping it fixed. We compare policy-aware task agent self-regulation with a trained policy guard. On PolicyGym, the final guard-based configuration achieves a 40-44% relative reduction in attack success rate and fewer benign rejections than policy prompting alone.

Read in: English · Spanish · Chinese · Hindi · Arabic · Portuguese · French · German · Japanese · Russian · Korean · Indonesian · Vietnamese

Topics: Executable, Guard (computer science), Computer science, Construct (python library), Enforcement, Task (project management)

Sources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange