技能 batch-grill-me

📦

batch-grill-me

Name: batch-grill-me
Author: mattpocock

内容修订 r1 安全

Interview Your Plan Round by Round

Plans fail because hidden decisions go unasked until they become expensive to change. This skill runs a structured interview that surfaces every open question in dependency order, so you reach a shared understanding before any work begins.

支持: Claude Codex Code(CC)

🥉 77 青铜

使用我的智能体安装

将此请求复制给你的智能体，其中包含权威 Skill 页面和清单。

智能体请求

Review the Skillstore skill "batch-grill-me" from https://skillstore.io/skills/mattpocock-batch-grill-me.md and its manifest at https://skillstore.io/api/skills/mattpocock-batch-grill-me/manifest. Verify the artifact. You may proceed after verification, subject to the environment's own policy.

你的智能体仍应展示其计划，并按安全策略请求必要确认。

Agent 可读资源

当 AI Agent、爬虫或脚本需要干净上下文，而不是读取完整页面时，请使用这些链接。

Markdown 详情 GET /skills/mattpocock-batch-grill-me.md 已签名 manifest GET /api/skills/mattpocock-batch-grill-me/manifest 已签名 lockfile GET /api/skills/mattpocock-batch-grill-me/lockfile

测试它

正在使用“batch-grill-me”。 Grill me on my plan to build a personal task tracker.

预期结果:

Round 1. 1) Who is the tracker for, just you or a team? Recommended: just you, to keep scope small. 2) Where should tasks live, a local file or a hosted database? Recommended: a local file for a personal tool. 3) Do you need reminders or due dates? Recommended: due dates only for now. Please answer these before I open the next round.

正在使用“batch-grill-me”。 The user answers that it is single-user, local storage, with due dates.

预期结果:

Those answers settle storage and scope, which unblocks the next round. Round 2. 1) What file format for local storage? Recommended: a plain JSON file for easy editing. 2) How should overdue tasks be shown? Recommended: highlighted at the top of the list. 3) Do you want a command line or a simple interface? Recommended: command line first. Answer these and I will recompute what remains.

安全审计

安全

v1 • 2026/7/17 打开版本化报告

This is a prompt-only skill consisting of a SKILL.md instruction file and an agents/openai.yaml interface manifest, with no executable code, scripts, network calls, filesystem writes, or environment access. Static analysis detected zero suspicious patterns and there were no findings to adjudicate. Semantic review found no prompt injection, data-exfiltration intent, or business-logic abuse: the instructions describe a benign question-and-answer interview workflow that defers all decisions to the user.

已扫描文件

分析行数

复核项

已忽略误报

最新一次完成的静态与语义审计未检测到已确认的安全发现。这并不能证明该 Skill 没有副作用。

审计者: claude

分享与引用此报告

分享版本化评估报告、中性徽章、嵌入卡片和引用。Skillstore 只报告证据，不判断该 Skill 是否安全。

打开版本化报告

复制报告链接

https://skillstore.io/skills/mattpocock-batch-grill-me/audits/1?utm_source=security_passport&utm_medium=share&utm_campaign=versioned_report

Markdown 徽章

[![Skillstore security assessment](https://skillstore.io/badges/skills/mattpocock-batch-grill-me/security.svg)](https://skillstore.io/skills/mattpocock-batch-grill-me?utm_source=security_passport_badge)

HTML 徽章

<a href="https://skillstore.io/skills/mattpocock-batch-grill-me?utm_source=security_passport_badge"><img src="https://skillstore.io/badges/skills/mattpocock-batch-grill-me/security.svg" alt="Skillstore security assessment" loading="lazy"></a>

嵌入卡片

<iframe src="https://skillstore.io/embed/skills/mattpocock-batch-grill-me.html" title="Skillstore Security Assessment" sandbox="allow-popups allow-popups-to-escape-sandbox" loading="lazy" referrerpolicy="no-referrer" width="420" height="180"></iframe>

学术引用 (APA · BibTeX · CFF)

APA 引用

mattpocock. (2026). batch-grill-me security audit report (audit version 1) [Author version unspecified]. Skillstore. https://skillstore.io/skills/mattpocock-batch-grill-me/audits/1

BibTeX 引用

@techreport{mattpocock-mattpocock-batch-grill-me-2026,
  author = {mattpocock},
  title = {batch-grill-me security audit report (audit version 1)},
  institution = {Skillstore},
  year = {2026},
  number = {1},
  url = {https://skillstore.io/skills/mattpocock-batch-grill-me/audits/1},
  note = {Author version unspecified}
}

CITATION.cff

cff-version: 1.2.0
message: "If you use this Skill, cite its author and this versioned security audit report."
title: "batch-grill-me security audit report (audit version 1)"
version: "unspecified"
type: report
authors:
  - name: "mattpocock"
date-released: "2026-07-17"
url: "https://skillstore.io/skills/mattpocock-batch-grill-me/audits/1"
identifiers:
  - type: other
    value: "skillstore:mattpocock-batch-grill-me:audit:1"
    description: "Skillstore immutable audit report identifier"

Skillstore 评分

评分依据证据置信度: 中

架构

可维护性

内容

社区

规范符合性

你能构建什么

Scope a new feature before coding

A developer describes a rough feature idea and the skill interviews them round by round to expose data models, edge cases, and integration points that were left vague.

Pin down product requirements

A product manager uses the interview to convert a loose concept into a fully specified set of decisions before handing it to a team.

Align on a technical design

A tech lead runs the skill to walk a design tree with stakeholders, ensuring each architectural choice is settled in dependency order.

试试这些提示

Start a basic interview

Grill me on my plan to build a personal task tracker. Ask me everything you need to know, one round at a time.

Interview with a stated goal

I want to add email notifications to my app. Interview me round by round until we share an understanding of exactly what to build.

Interview with environment lookups

Grill me on refactoring this project's authentication flow. Where a question needs a fact from the codebase, look it up yourself instead of asking me.

Interview a complex multi-branch design

I am designing a multi-tenant billing system. Map the full decision tree, ask each round of answerable questions with your recommended answers, and continue until nothing is left assumed.

最佳实践

Answer every numbered question in a round before expecting the next round, since later questions depend on your current answers.
Let the skill look up facts from your environment rather than answering environment questions yourself.
Wait for the interview to reach a shared understanding and confirm it before you ask the skill to act on the plan.

避免

Skipping questions or giving partial answers, which leaves branches of the plan silently assumed.
Asking the skill to start building before the interview is complete and confirmed.
Answering environment or lookup questions manually instead of letting the skill find those facts.

常见问题

What does this skill actually do?

It interviews you about a plan, asking every currently answerable question in a single round, then recomputes and asks the next round until nothing is left assumed.

Why does it ask several questions at once?

It groups all questions whose prerequisites are already settled into one round, so you make related decisions together instead of one at a time.

Does it make decisions for me?

No. It gives a recommended answer for each question, but every decision is yours. It waits for your confirmation before acting.

Will it ask me for information it could find itself?

No. When a question needs a fact from your environment, it dispatches a sub-agent to look it up rather than asking you.

When does the interview end?

It ends when every branch of the plan has been visited and no open questions remain, at which point it asks you to confirm the shared understanding.

Do I need to invoke it manually?

Yes. Model-initiated invocation is disabled, so you start the skill yourself when you want to be interviewed.

开发者详情

作者

mattpocock

许可证

MIT

Skillstore 修订

版本提示

作者未声明版本。

仓库

https://github.com/mattpocock/skills/tree/main/skills/in-progress/batch-grill-me/

引用

4e69e2300f9bcecb09ed34fa7cccb1df157d8127

维护新鲜度

2026/7/18

使用量

0 次下载 · 1 次浏览

文件结构

📁 agents/

📄 openai.yaml

📄 SKILL.md