智能体技能安全研究

Versioned Transitive Dependency-Closure Binding and Operation-Time Effect Governance for Agent Skills: ClosureBound

面向智能体技能的版本化传递依赖闭包绑定与操作时效果治理:ClosureBound

签名后的技能目录并不等于完整的运行中技能。ClosureBound 将权限绑定到完整依赖闭包,并在外部效果持久化时再次核验。

黑色像素马赛克依赖分支汇聚为绑定闭包,穿过操作时检查边界后,仅保留一条受限效果路径。

研究概览

一个智能体技能可能依赖签名根目录之外的文件、软件包、工具、模型与远程服务。这些依赖可能递归变化,或只在运行时延迟加载,使实际运行系统已经发生漂移,而根目录层面的证据看起来仍然有效。

ClosureBound 将这种异构依赖图解析为类型化闭包,把每项授权绑定到精确闭包根与效果上限,并在操作时重新解析闭包和状态。只有一个联合见证同时满足全部边界,持久化效果才可继续;通往相同效果的替代路径也不能放大权限。

核心贡献

  1. 01

    运行身份定义

    不再仅凭 manifest 或签名目录识别智能体技能,而是使用其类型化传递依赖闭包定义实际运行身份。

  2. 02

    操作时联合见证

    在效果持久化时重新解析闭包与状态,对等价操作入口进行规范化,并要求一个见证同时满足全部约束。

  3. 03

    机械化不变量验证

    形式化版本不继承、效果不放大、新鲜度与路径不变性,并通过生命周期夹具、内核契约、变异体和状态探索进行检验。

  4. 04

    公开技能审计

    审计 549 个公开 Skills,衡量打包依赖被显式声明的稀缺程度,以及路径解析到技能根目录之外的情况。

ClosureBound 如何扩展治理技术栈

OpenPort、IGAC 与 EBTE 管理哪些工具效果具备执行资格及其依据。ClosureBound 增加了一个互补问题:正在执行的技能,是否仍是获得授权时那个依赖闭合的系统?

授权会绑定到已解析闭包、目的、来源、有效期、epoch 与效果上限。任何重要依赖或状态变化,都必须在持久化效果发生前反映到新的操作时见证中。

  1. 01智能体技能
  2. 02已解析闭包
  3. 03绑定授权
  4. 04受治理效果

证据与适用范围

本文目前是 arXiv 预印本与参考监控器设计,并未宣称已具备生产就绪性。其保证依赖完整中介与依赖发现、经过认证的新鲜度、正确的效果规范化、密码学绑定和权威线性化。论文不证明程序等价性或远程服务诚实性;公开技能研究属于词法审计,也不代表可利用风险的实际发生率。

英文摘要

Agent Skills combine instructions with files, packages, tools, models, and services, so operational identity can exceed a signed directory. Recursive or lazy dependencies may change while root-level evidence remains valid, and different surfaces may reach the same durable effect. We present ClosureBound, a reference monitor that prevents authorization transfer across material changes to this heterogeneous closure. Its resolver commits typed graph nodes and topology. Each grant binds an exact closure root, effect ceiling, purpose/provenance, validity, and epochs. At durability, it re-resolves closure and state, normalizes the operation into an external-effect IR, and admits it only if a joint witness satisfies every bound. Supported equivalent paths share one ceiling. Assuming complete mediation and discovery, authenticated freshness, sound normalization, cryptographic binding, and authoritative linearization, we establish metadata non-authority, closure determinism, version non-inheritance, effect non-amplification, bound-value freshness, and path invariance. We do not establish program equivalence or remote-service honesty. A provider-free implementation matches 40 frozen lifecycle fixtures; 18 kernel contracts and six mutants cover binding and downgrade cases. Full-profile exploration reaches 84,608 states and 530,752 transitions without a declared invariant violation; six weakened profiles yield witnesses. A lexical audit of 549 public Skills (4,872 unique files) finds that 21 of 526 roots with bundled files name every non-manifest path verbatim, 67 contain links resolving outside their roots, and no root declares a frontmatter dependencies field. These observations motivate conservative closure discovery and define concrete targets for broader runtime, interoperability, efficacy, and production validation.

引用本文

请通过 DOI 记录获取稳定的书目信息和引用格式。