菜单
Grill the user relentlessly about a plan, decision, or idea — a lightweight interview that writes no docs. Use for stress-testing thinking inline, or on any 'grill' trigger; for a feature grill that also writes CONTEXT.md and ADRs, use /grill-me.
默认使用会先检查来源的 Prompt；你也可以切换为直接命令，或下载本地副本。
决定是否安装前，请先阅读 SKILL.md，以及 SkillsMP 当前展示的配套文件。
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。
直接命令不会经过审查 Prompt；运行前请先检查来源。
npx skills add https://github.com/scheron/dotfiles --skill grill命令会保持在同一行。复制前请横向滚动并检查完整内容。
想先保存到本地？可下载 SkillsMP 当前能够提供的文件。
Daily is the user's personal task tracker; the `daily` CLI reads and changes its tasks. Use when the user names Daily about tasks (also spelled "дейли"), asks about their own day or to-dos without naming it ("what do I have today", "what was I doing on Monday", "что у меня сегодня"), or wants a task captured, completed, rescheduled, searched, estimated or time-logged.
Столкнуть Claude Code и Codex — две модели разных провайдеров — в споре по одной теме, до согласованного решения или эскалации человеку. Use when the user wants a second opinion from a different provider, a cross-model debate, or to stress-test a design decision against an independent model. Not for solo ideation.
Russian ultra-compressed communication mode based on JuliusBrussee/caveman PR #144. Cuts token usage by answering in terse Russian while keeping full technical accuracy. Supports ru-lite, ru-full default, ru-ultra, and ru-notes. Use when user says caveman-ru, /caveman-ru, /caveman ru, пещерный режим, режим пещерного, меньше токенов, короче, кратко, or asks for Russian token-efficient replies.
基于 SOC 职业分类
正在显示 SKILL.md
| name | grill |
| description | Grill the user relentlessly about a plan, decision, or idea — a lightweight interview that writes no docs. Use for stress-testing thinking inline, or on any 'grill' trigger; for a feature grill that also writes CONTEXT.md and ADRs, use /grill-me. |
Interview me relentlessly about every aspect of this until we reach a shared understanding. Walk down each branch of the decision tree, resolving dependencies between decisions one-by-one. For each question, provide your recommended answer.
Ask the questions one at a time, waiting for feedback on each question before continuing. Asking multiple questions at once is bewildering.
If a fact can be found by exploring the environment (filesystem, tools, etc.), look it up rather than asking me. The decisions, though, are mine — put each one to me and wait for my answer.
Where a question has genuine alternatives, put 2-3 of them on the table — each with its trade-offs — then lead with the one you recommend and say why. A single recommended answer walks us down one branch of the tree without ever showing that the others existed; that is how a plan comes out solid and still wrong. Be opinionated: I want a strong read, not a menu.
Do not act on it until I confirm we have reached a shared understanding.