菜单
Use for safe OpenShell policy-boundary demos and denied-operation analysis.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
| name | openshell-policy |
| description | Use for safe OpenShell policy-boundary demos and denied-operation analysis. |
Use this skill when the participant asks to demonstrate sandbox boundaries, network restrictions, filesystem restrictions, or policy enforcement.
Ask the agent to attempt one blocked action, stop, and explain the failure. Do not ask it to find another route around the block.
For the workshop fixture, writing under /sandbox should be allowed and writing under protected system paths such as /etc should be denied unless the instructor changes the policy.
Analyze the workshop application log from inside OpenShell and produce incident artifacts.
Interpret traces from a parent Deep Agent and subagents running inside OpenShell.
Demonstrate and explain the contained agent runtime's filesystem and network boundaries.
Use for analyzing application logs with Python inside the OpenShell sandbox and producing incident reports.
Use for interpreting LangSmith traces, adding useful metadata, and comparing agent runs.
