跳转到正文
报告库
用途分类 / 其他用途

Grilling Skill 安全审计

作者说它能做什么(原文)

Grill the user relentlessly about a plan, decision, or idea. Use when the user wants to stress-test their thinking, or uses any 'grill' trigger phrases.

第三方安全检查结论

发现安全风险

已检查文件
2
发现的风险
3
会不会运行危险命令?检查是否下载程序后直接运行、让他人远程控制电脑,或藏起要运行的命令。未发现风险
会不会泄露文件和密钥?检查是否发送含密码或密钥的文件,以及代码里是否直接写了密钥。发现 1 项风险
中风险

环境查询和自动派遣子代理可能扩大敏感文件的接触范围

原文依据:1 处
发现了什么

Skill 要求凡是能从环境查到的事实都自行查找,并明确要求派遣子代理,但没有限定允许读取的目录、文件类型、敏感信息范围,也没有要求查询前取得用户同意。

为什么需要注意

如果待讨论的问题可由工作区、配置文件或工具状态回答,代理和子代理可能读取与决策无关或含凭据、客户资料、商业信息的内容。所示文件没有要求将数据发送到外部,但额外代理会扩大接触这些内容的主体和上下文。

这段证据能说明什么

第26行确实要求代理自行查询环境事实并派遣子代理,且未在 Skill 内规定目录或敏感文件边界,因此可能增加环境数据被读取和在代理间传递的范围。但它只授权查找当前问题所需的“事实”,没有要求读取凭据、搜遍文件系统或向外部发送数据;实际风险取决于运行环境权限及任务内容。用户可要求作者明确查询范围、排除敏感路径并在读取敏感资料前确认。

这项判断针对展示的代码和适用条件,不表示风险已经实际发生。
SKILL.md:26来自说明文档打开原文件
Finding _facts_ is your job, never the user's. When a frontier question needs a fact from the environment (filesystem, tools, etc.), dispatch a sub-agent to find it; don't ask the user for anything you could look up yourself. Don't block on it: a running exploration is an unsettled prerequisite, so only the questions downstream of it wait for the sub-agent to report; ask the rest of the frontier now. The _decisions_ are the user's: put each to them and wait.
会不会删除文件或一直在后台运行?检查是否大范围删除文件、改写磁盘,或设置自动启动。未发现风险
会不会绕过安全保护?检查是否跳过网站安全验证、开放过多文件权限,或取消操作前的确认。发现 1 项风险
中风险

无条件要求派遣子代理会绕过用户对委派范围的逐次控制

原文依据:2 处
发现了什么

只要需要环境事实,指令就要求派遣子代理,没有把委派设为可选,也没有要求说明子代理将访问什么或先征得批准。

为什么需要注意

在允许子代理继承工具或工作区权限的平台上,一次普通的规划讨论可能自动扩展为额外代理活动,使用户更难掌握谁在检查哪些资源。

第26行把派遣子代理写成必须执行的步骤,没有要求事先说明委派范围或逐次取得用户许可,因此用户可能无法在每次委派前控制哪些任务和环境事实被交给另一代理。风险限于查找当前前置事实,且源码未授予额外权限或外部通信;实际访问仍受宿主权限约束。用户可要求默认不委派,或在委派前列出目标、数据范围和权限并征得同意。

SKILL.md:26来自说明文档打开原文件
Finding _facts_ is your job, never the user's. When a frontier question needs a fact from the environment (filesystem, tools, etc.), dispatch a sub-agent to find it; don't ask the user for anything you could look up yourself. Don't block on it: a running exploration is an unsettled prerequisite, so only the questions downstream of it wait for the sub-agent to report; ask the rest of the frontier now. The _decisions_ are the user's: put each to them and wait.
查看另外 1 个位置
SKILL.md:28来自说明文档打开原文件
The session is done when the frontier is empty: every branch of the design tree visited, nothing left silently assumed. Do not act on it until the user confirms you have reached a shared understanding.
会不会误导 AI 或隐藏内容?检查工作说明是否要求 AI 忽略你的指令、干扰检查结果,或夹带看不见的文字。发现 1 项风险
中风险

“无休止”提问并为每题推荐答案可能造成疲劳和决策锚定

原文依据:3 处
发现了什么

Skill 两次要求持续、彻底地盘问,并要求每个问题都附上自己的推荐答案;结束条件是遍历所有分支,而不是用户认为信息已经足够或选择停止。

为什么需要注意

在复杂计划中,用户可能面对大量连续问题,并因疲劳而接受预先推荐的选项。推荐措辞还可能把用户的商业、权限或风险决策锚定到代理偏好的答案上。

这是活跃的交互规则:Skill 要求“relentlessly”访谈、每轮为每题给推荐答案,并把完成条件设为所有分支均已遍历。反复暴露于推荐答案可能锚定用户判断,而“frontier 为空”才结束可能造成疲劳。不过它也要求等待用户回答且行动前取得确认,因此并非替用户执行决定。用户可要求允许随时停止、限制轮数,并在其先回答后再显示建议。

SKILL.md:6来自说明文档打开原文件
Interview the user relentlessly until you reach a shared understanding. Map this as a **design tree**: every decision branches into the decisions that hang off it.Work the tree in **rounds**. The **frontier** is every decision whose prerequisites are already settled: the questions you can ask _now_ without guessing at answers you haven't heard yet. Ask the whole frontier in one round: number each question and give your recommended answer. Then wait for the user's answers before the next round.
查看另外 2 个位置
SKILL.md:28来自说明文档打开原文件
The session is done when the frontier is empty: every branch of the design tree visited, nothing left silently assumed. Do not act on it until the user confirms you have reached a shared understanding.
SKILL.md:24来自说明文档打开原文件
Each round the user answers reshapes the tree: settled decisions push the frontier outward and unblock questions that depended on them. Recompute the frontier and ask the next round. A question whose answer depends on another question still open in this round belongs to a _later_ round, not this one.
会不会偷偷改推广链接或收款方?检查是否强制替换推广链接或收款对象,同时要求隐瞒更改。未发现风险

Skill 逻辑拆解

0 个说明模块

该 Skill 的主要行为是把用户的计划或决定拆成决策树,并分轮询问当前可回答的全部问题;每个问题都附带 Skill 自己推荐的答案。

查看原文
SKILL.md:6来自说明文档打开原文件
Interview the user relentlessly until you reach a shared understanding. Map this as a **design tree**: every decision branches into the decisions that hang off it.Work the tree in **rounds**. The **frontier** is every decision whose prerequisites are already settled: the questions you can ask _now_ without guessing at answers you haven't heard yet. Ask the whole frontier in one round: number each question and give your recommended answer. Then wait for the user's answers before the next round.

当问题需要环境事实时,Skill 指示代理自行查询并派遣子代理,而不是询问用户;下游问题等待查询结果,其他问题继续进行。

查看原文
SKILL.md:26来自说明文档打开原文件
Finding _facts_ is your job, never the user's. When a frontier question needs a fact from the environment (filesystem, tools, etc.), dispatch a sub-agent to find it; don't ask the user for anything you could look up yourself. Don't block on it: a running exploration is an unsettled prerequisite, so only the questions downstream of it wait for the sub-agent to report; ask the rest of the frontier now. The _decisions_ are the user's: put each to them and wait.

Skill 明确要求在所有决策分支处理完并由用户确认共同理解前,不执行该计划。

查看原文
SKILL.md:28来自说明文档打开原文件
The session is done when the frontier is empty: every branch of the design tree visited, nothing left silently assumed. Do not act on it until the user confirms you have reached a shared understanding.
从这里开始 · 工作说明SKILL.md
grilling

工作说明没有分章节。可在下方查看随附文件和相关代码。

连线表示工作说明包含的模块,不是实际运行顺序。点击模块可查看原文。
文件与检查记录2 个文件

检查范围与遗漏

逐文件查看涉及的内容

下方列出本次涉及的原文范围;纳入检查不代表已查清所有问题。

  • SKILL.md已纳入全文
  • agents/openai.yaml已纳入全文

这份报告只针对上方版本。我们看了拿到的代码和说明文件,没有实际运行 Skill,也没有检查它另外安装的软件包。因此,这不是“保证安全”的承诺;换了版本或使用环境,结果也可能不同。

  • SKILL.md工作说明
  • agents/openai.yaml配套文件
读取了多少行
33
文件校验值(用于核对版本)
7c1aab9fb5cdec80183358bbf79c4fd5242e5d210d3cf48e2c1ca5436deb92d6