4x4
Pull 16 fast decisions out of the user with 4 sets of 4 open questions, then present an implementation plan. Only for risky changes, large product decisions, or large refactors. Run when the user types /4x4. Agents may suggest it in one line but never start it on their own.
- License 许可
- MIT
- 最近更新
- 2026-11-14
- Stars
- ★ 4.3k
- Forks
- 626
01 前言 Front Matter
- name
- 4x4
- description
- Pull 16 fast decisions out of the user with 4 sets of 4 open questions, then present an implementation plan. Only for risky changes, large product decisions, or large refactors. Run when the user types /4x4. Agents may suggest it in one line but never start it on their own.
03 SKILL.md 正文
4x4
Get the user's judgment, taste, and decisions in 4 quick rounds. 16 open questions. A full run should take them 2–3 minutes.
The 4 sets
These are defaults. Adapt their names and focus to pull the most judgment and taste out of the user for this topic.
- Product — what it does, who it's for, UI/UX, scope.
- Software design — architecture, data, boundaries, what to reuse or delete.
- Practical — execution order, rollout, what ships first.
- Edge cases — the 4 most likely ways this breaks or surprises users. Each question asks how to handle one.
Workflow
- Load context fast. Use what's already in the thread. Skim only the files and docs that matter (scopes, ADRs, the diff). No research, no subagents.
- Pick the order. Start with the set whose answers most change the others.
- Pick 4 questions per set. Choose the 4 decisions with the biggest impact that only the user can make. Never ask what the code, docs, or common sense already answer. Skip anything already decided.
- Show one set at a time. Never preview or list the other sets. Stop and wait. Show the next set only after the user answers.
- Adapt. Rewrite later sets based on earlier answers.
- Present the plan. After set 4, give plain step-by-step instructions to execute the work properly, built from their answers. Chat only. Then stop. Don't build or write docs until the user says go.
Message format
## Set 2/4 · Software design
1. Where should X live?
2. ...
3. ...
4. ...
Nothing else. No intro, no outro.
Rules
- Open questions only. Never give options, picks, recommendations, or hints. The point is the user's own thinking.
- One short sentence per question. Concrete, plain English, clear on first read.
- No meta questions, like how to measure success or what to test first. Ask about the work itself.
- Vague or skipped answer: move on and use your judgment.
- If the user calls a question bad, replace it or drop it. Your call.
- If an answer contradicts an ADR or scope, point it out in the plan.
- If an answer opens a big new question, note it in the plan. Don't add a 5th question.
04 同仓库更多 Skills
varied-search
Research a topic across the web, GitHub, and X/Twitter with DeepAPI and produce a concise Markdown report. Use for varied-search or requests to compare web coverage, relevant repositories, and social discussion.
total-review
Run five independent max-effort reviews (Fable 5.1, Opus 5.5, GPT-6 Astra, GPT-6.1 Sol, Muse Wonder), merge their findings, and fix only what the user approves. Use for /total-review, or when the user or another agent asks for a total review.
stop-overthinking
Manually-invoked skill that stops overthinking and forces a short, practical decision. Use only when the user explicitly invokes stop-overthinking.
state
Manually-invoked skill that recaps what the current thread is doing and how far along it is. Use when the user says "state", "where are we", or "what % are we done".
setup-help
Guide setup one step at a time and show what remains. Use only when the user explicitly asks for it by name.
self-archive
Archive the current Cloudroom thread and release its runtime. Use only when the user explicitly asks to self-archive, archive yourself, or archive this thread.