今日速览
| 序号 | 标题 | 来源 | 日期 | 主题 | 推荐等级 |
|---|---|---|---|---|---|
| 1 | Adaptive Skill Granularity for Coding Agents: Runtime Routing and Macro-to-Micro Recovery | arXiv | 2026-10-22 | AI-Agent | 高 |
| 2 | Certified Execution Harness for Agentic AI: Safe and Least-Intrusive Petri-Net Supervisory Control Between Reasoning and Action | arXiv | 2026-10-22 | AI-Agent | 高 |
重点论文与技术动态
1. Adaptive Skill Granularity for Coding Agents: Runtime Routing and Macro-to-Micro Recovery
- 来源：arXiv
- 日期：2026-10-22
- 作者/机构：Dingxuan Zhang, Yanhui Jiang, Jun Wang, Hefeng Zhou
- 主题标签：
AI-Agent,arXiv - 推荐等级：高
- 分类：cs.AI, cs.SE
一句话结论
论文提出自适应技能粒度框架，通过运行时路由与“宏观到微观”失败恢复，解决了代码智能体固定技能粒度的脆弱性问题。
核心内容
- 指出固定宏技能高效但脆弱，宏到微的混合执行能解决单粒度无法应对的失败。
- 将技能粒度构建为依赖任务结构、执行状态和失败反馈的动作抽象策略。
- 强调技能库不仅是检索资源，更是可控、可验证与可恢复的执行抽象。
方法与数据
- 结合特征粒度控制器与有限状态恢复执行器；基于确定性协议与多维验证指标评估，具体数据集摘要未明确。
价值判断
- 值得关注：从预执行检索拓展至运行时动态控制，证明显式降级恢复能显著增强智能体鲁棒性。
- 可复用点：特征驱动的粒度控制器架构与有限状态恢复执行器设计。
- 局限/待核查：具体评估所用的代码基准数据集与额外计算开销摘要未明确。
摘要
Reusable skills allow large language model agents to package task-specific procedures, tool affordances, and execution guidance into modular units. Prior work has largely studied how agents retrieve, compose, or synthesize skills before execution. In coding agents, however, execution poses an additional runtime control problem: the same task may be solved through primitive tool calls, macro skills, or hybrid execution that starts with a macro and falls back to micro-level recovery. We introduce Adaptive Skill Granularity, a framework that treats skill granularity as an action-abstraction policy conditioned on task structure, execution state, and macro-skill failures. The system combines a feature-based granularity controller and a finite-state recovery executor. We evaluate the framework with a deterministic coding-agent protocol using verifier-defined success, total cost, and process metrics such as macro adoption, failure recovery, route match, and transition correctness. Results show that fixed macro execution is efficient when applicable but brittle under mismatch, while runtime routing and explicit macro-to-micro recovery improve robustness and recover controlled macro failures that fixed-granularity execution leaves unresolved. Our findings suggest that skill libraries for coding agents should be treated not only as retrievable resources, but also as controllable, verifiable, and recoverable execution abstractions.2. Certified Execution Harness for Agentic AI: Safe and Least-Intrusive Petri-Net Supervisory Control Between Reasoning and Action
- 来源：arXiv
- 日期：2026-10-22
- 作者/机构：Chao Gu, Quanyan Zhu
- 主题标签：
AI-Agent,arXiv - 推荐等级：高
- 分类：cs.AI, eess.SY
一句话结论
本文提出基于 Petri 网监督控制的智能体执行框架，在提示注入攻击下提供形式化安全保证，并优化了带重试成本的闭环重规划侵入性。
核心内容
- 形式化防御：以 Petri 网标记代表能力、审批与预算；只要前置条件未满足即拒绝执行，不受提示词措辞影响。
- 侵入性反常现象：引入重试消耗推理成本的重规划闭环，证明传统的“最大许可监督器”可能比更严格的监督器更具侵入性。
- 最小侵入优化：提出基于价值阈值细化的最小侵入监督器，并在自主发布智能体上证实安全部署需包含部署启动后仍可拒绝的步骤。
方法与数据
- 采用离散事件系统中的 Petri 网监督控制理论进行建模与综合；在自主发布智能体上验证，具体实验数据集摘要未明确。
价值判断
- 值得关注：运用形式化控制理论阻断提示注入攻击，为智能体工具调用提供确定性安全保障。
- 可复用点：利用 Petri 网标记管理权限、配额及并发动作约束的设计范式。
- 局限/待核查：运行时的安全性传递依赖4项特定假设，在复杂实际环境中的满足程度有待核查。