vibe-coding

Compare original and translation side by side

🇺🇸

Original

English
🇨🇳

Translation

Chinese

Vibe Coding

Vibe Coding

Scope

范围

Covers
  • Timeboxed, AI-assisted rapid prototyping (“vibe coding”) to produce a functional demo (not slides)
  • Turning a rough idea into a buildable prototype spec + task board + prompt pack
  • A tight iteration loop: generate → run → verify → adjust → log decisions
  • “Build tools to build the thing” when it meaningfully speeds up the demo (timeboxed)
  • Safe use of coding agents: least privilege, no secrets, small diffs, validation, rollback
When to use
  • “Vibe code a working prototype we can demo in 30–90 minutes.”
  • “Replace this Figma concept with a clickable prototype.”
  • “I’m not an engineer—help me build a small app/tool with AI and ship a demo.”
  • “Turn this AI feature idea into a proof-of-concept with a clear build loop and demo script.”
When NOT to use
  • You need a production-grade system, hardening, scaling, or security review (use
    building-with-llms
    + engineering process).
  • You need upstream problem framing, strategy, or PRD-level alignment (use
    problem-definition
    ,
    writing-prds
    ).
  • The work is high-stakes/irreversible (payments, auth, medical, legal, safety-critical) without human owners and reviews.
  • The request is “build anything” with no demo target; do intake first and narrow to one scenario.
涵盖内容
  • 限时、AI辅助的快速原型开发(“Vibe Coding”),产出可运行的演示原型(非幻灯片形式)
  • 将粗略想法转化为可落地的原型规格+任务看板+提示包
  • 紧凑的迭代循环:生成→运行→验证→调整→记录决策
  • 当能显著加快演示进度时,“构建工具来实现目标”(限时进行)
  • 安全使用编码Agent:最小权限、无敏感信息、小幅度变更、验证机制、回滚方案
适用场景
  • “通过Vibe Coding在30–90分钟内开发出可演示的工作原型。”
  • “将Figma概念原型替换为可点击的交互式原型。”
  • “我不是工程师——帮我用AI构建一个小型应用/工具并完成演示版本交付。”
  • “将这个AI功能想法转化为带有清晰构建循环和演示脚本的概念验证原型。”
不适用场景
  • 你需要生产级系统、加固、扩容或安全评审(请使用
    building-with-llms
    +工程流程)。
  • 你需要上游问题梳理、战略规划或PRD级别的对齐(请使用
    problem-definition
    writing-prds
    )。
  • 工作涉及高风险/不可逆转的场景(支付、认证、医疗、法律、安全关键领域)且无人工负责人和评审。
  • 需求为“随便做个什么”且无明确演示目标;请先进行需求收集并缩小到单一场景。

Inputs

输入信息

Minimum required
  • Prototype goal: what should a user be able to do in the demo (1–3 “happy path” tasks)
  • Target user + context (who uses it, where it fits)
  • Timebox (e.g., 30/60/90 minutes) + demo audience (internal, customer, exec)
  • Platform preference (web app, mobile, CLI, spreadsheet, etc.) and constraints (privacy, data sensitivity)
  • Data/integrations: mock data ok? any APIs needed? (default to mock)
Missing-info strategy
  • Ask up to 5 questions from references/INTAKE.md (3–5 at a time).
  • If details remain missing, proceed with explicit assumptions and offer 2–3 options (e.g., mock vs real data; simple UI vs polished).
  • If asked to run commands or write/modify files, request confirmation, keep changes in a dedicated folder, and include rollback guidance.
必填输入信息
  • 原型目标:演示中用户应能完成的操作(1–3条“顺畅路径”任务)
  • 目标用户+使用场景(谁使用、适用场景)
  • 时间限制(如30/60/90分钟)+演示受众(内部团队、客户、高管)
  • 平台偏好(Web应用、移动端、CLI、电子表格等)及约束条件(隐私、数据敏感性)
  • 数据/集成:是否允许使用模拟数据?是否需要调用API?(默认使用模拟数据)
信息缺失处理策略
  • references/INTAKE.md中最多提出5个问题(每次3–5个)。
  • 若仍有信息缺失,基于明确假设推进工作,并提供2–3种选项(如模拟数据vs真实数据;简易UI vs 精致UI)。
  • 若要求执行命令或编写/修改文件,需先请求确认,将变更放在专用文件夹中,并提供回滚指导。

Outputs (deliverables)

交付成果

Produce a Vibe Coding Prototype Pack (in chat; or as files if requested), in this order:
  1. Vibe Coding Brief (goal, demo scenario, non-goals, constraints, timebox)
  2. Prototype Spec (user flow, screens/components, data model, acceptance criteria, “fake vs real” decisions)
  3. Prompt Pack (copy/paste prompts to drive the coding agent safely and efficiently)
  4. Build Plan + Task Board (vertical slices with checks/tests per slice)
  5. Demo Script + Runbook (how to run, how to demo, what to say, what to avoid)
  6. Risks / Open questions / Next steps (always included)
Templates: references/TEMPLATES.md
产出Vibe Coding原型包(在对话中交付;若有需求可提供文件),顺序如下:
  1. Vibe Coding 简要说明(目标、演示场景、非目标、约束条件、时间限制)
  2. 原型规格(用户流程、界面/组件、数据模型、验收标准、“模拟vs真实”决策)
  3. 提示包(可直接复制粘贴的提示词,用于安全高效地驱动编码Agent)
  4. 构建计划+任务看板(按垂直切片划分,每个切片包含检查/测试环节)
  5. 演示脚本+操作手册(运行方式、演示方式、讲解内容、避坑要点)
  6. 风险/待解决问题/下一步计划(必须包含)
模板参考:references/TEMPLATES.md

Workflow (7 steps)

工作流程(7个步骤)

1) Pick a single demo outcome (kill ambiguity fast)

1) 确定单一演示成果(快速消除歧义)

  • Inputs: Initial idea, timebox, target audience.
  • Actions: Write a one-sentence demo promise (“In 60 minutes we will demo…”) + 3–5 non-goals. Choose one “hero” scenario and what can be faked.
  • Outputs: Draft Vibe Coding Brief.
  • Checks: The demo promise is specific, observable, and fits the timebox.
  • 输入信息:初始想法、时间限制、目标受众。
  • 行动:撰写一句演示承诺(“我们将在60分钟内完成……的演示”)+3–5条非目标。选择一个核心场景,并确定可模拟的内容。
  • 产出:草稿版Vibe Coding 简要说明
  • 校验:演示承诺需具体、可观测,且符合时间限制。

2) Define the prototype’s contract (what exists, what’s mocked)

2) 定义原型约定(哪些是真实功能,哪些是模拟内容)

  • Inputs: Demo scenario, platform preference, constraints.
  • Actions: Specify the minimum user flow, screens/components, and data shape. Decide: mock data vs real data; stub integrations vs live.
  • Outputs: Draft Prototype Spec.
  • Checks: Acceptance criteria exist for each user-visible step; “fake vs real” is explicit.
  • 输入信息:演示场景、平台偏好、约束条件。
  • 行动:明确最简用户流程、界面/组件及数据结构。决策:模拟数据vs真实数据;集成桩vs实时集成。
  • 产出:草稿版原型规格
  • 校验:每个用户可见步骤都有验收标准;“模拟vs真实”的决策清晰明确。

3) Set the build loop + guardrails (how we’ll vibe code safely)

3) 设置构建循环+防护规则(安全进行Vibe Coding)

  • Inputs: Repo/app context (if any), constraints, desired stack.
  • Actions: Create a Prompt Pack that forces: small diffs, clear file list, run instructions, and “ask before risky actions.” Create a task board of 3–8 vertical slices.
  • Outputs: Prompt Pack + Build Plan + Task Board.
  • Checks: Every slice has a Definition of Done and a quick validation method (manual steps or tests).
  • 输入信息:代码库/应用上下文(如有)、约束条件、技术栈偏好。
  • 行动:创建提示包,强制要求:小幅度变更、清晰的文件列表、运行说明、“风险操作前需确认”。创建包含3–8个垂直切片的任务看板。
  • 产出提示包+构建计划+任务看板
  • 校验:每个切片都有完成定义及快速验证方法(手动步骤或测试)。

4) Scaffold the thinnest runnable slice (end-to-end)

4) 搭建最简可运行切片(端到端)

  • Inputs: Prompt pack, chosen platform/stack, prototype spec.
  • Actions: Generate a minimal skeleton that runs. Implement the hero path with mock data. Capture run commands and known limitations.
  • Outputs: Runnable prototype + run notes (for the runbook).
  • Checks: A fresh user can run it in ≤ 5 minutes; the hero path is demonstrable.
  • 输入信息:提示包、选定的平台/技术栈、原型规格。
  • 行动:生成最小可运行骨架。基于模拟数据实现核心场景。记录运行命令及已知限制。
  • 产出:可运行原型+运行笔记(用于操作手册)。
  • 校验:新用户可在≤5分钟内启动运行;核心场景可正常演示。

5) Iterate in vertical slices (generate → run → verify → log)

5) 按垂直切片迭代(生成→运行→验证→记录)

  • Inputs: Task board, working prototype.
  • Actions: For each slice: request a plan + diff, apply changes, run, verify against acceptance criteria, and record decisions/bugs. Avoid broad refactors; prefer incremental improvements.
  • Outputs: Updated prototype + iteration notes.
  • Checks: Each slice ends with a user-visible improvement and a validated run.
  • 输入信息:任务看板、可运行原型。
  • 行动:针对每个切片:请求方案+变更内容,应用变更,运行验证,对照验收标准检查,记录决策/问题。避免大规模重构;优先选择增量改进。
  • 产出:更新后的原型+迭代笔记。
  • 校验:每个切片迭代后需实现用户可见的功能改进,并完成运行验证。

6) Optional: build a tool to build the thing (timeboxed)

6) 可选:构建辅助工具(限时)

  • Inputs: Repeated friction (editing, generating, transforming content).
  • Actions: If it reduces time-to-demo, vibe code a tiny helper tool (editor, generator, script) and immediately use it to advance the prototype.
  • Outputs: Helper tool + note on how it accelerates the workflow.
  • Checks: The helper tool saves time within the current timebox; otherwise cut it.
  • 输入信息:重复出现的痛点(编辑、生成、转换内容时的摩擦)。
  • 行动:若能缩短演示准备时间,可通过Vibe Coding开发一个小型辅助工具(编辑器、生成器、脚本),并立即用它推进原型开发。
  • 产出:辅助工具+关于其如何加速工作流程的说明。
  • 校验:辅助工具需在当前时间限制内节省时间;否则取消该步骤。

7) Package the demo + quality gate + handoff

7) 打包演示内容+质量校验+交接

  • Inputs: Prototype + all draft artifacts.
  • Actions: Write the demo script + runbook. Run references/CHECKLISTS.md and score with references/RUBRIC.md. Finalize Risks / Open questions / Next steps.
  • Outputs: Final Vibe Coding Prototype Pack.
  • Checks: A stakeholder can demo it without you; risks and next steps are explicit and owned.
  • 输入信息:原型+所有草稿文档。
  • 行动:编写演示脚本+操作手册。执行references/CHECKLISTS.md并通过references/RUBRIC.md评分。最终确定风险/待解决问题/下一步计划
  • 产出:最终版Vibe Coding原型包
  • 校验:相关人员无需你的协助即可完成演示;风险及下一步计划清晰明确且有负责人。

Quality gate (required)

质量校验(必填)

  • Use references/CHECKLISTS.md and references/RUBRIC.md.
  • Always include: Risks, Open questions, Next steps.
  • 使用references/CHECKLISTS.mdreferences/RUBRIC.md进行校验。
  • 必须包含:风险待解决问题下一步计划

Examples

示例

Example 1 (30–60 min prototype): “Use
vibe-coding
to build a demo-ready web prototype of an ‘AI meeting notes → action items’ tool. Mock the LLM output. Output the full Vibe Coding Prototype Pack.”
Expected: brief + spec + prompt pack + task board + demo script; prototype plan defaults to mock data and a single hero flow.
Example 2 (non-engineer builder): “I’m a PM. Use
vibe-coding
to help me create a clickable prototype of an onboarding checklist app in 45 minutes. I need a demo script for my team.”
Expected: tight scope, fake data, vertical slices, and a runbook optimized for demo reliability.
Boundary example: “Vibe code a production payments backend and deploy it.”
Response: out of scope; propose a prototype-only approach (mock payments), identify required security/engineering owners, and recommend a separate production plan.
示例1(30–60分钟原型):“使用
vibe-coding
开发一个可演示的Web原型,实现‘AI会议纪要→行动项’工具。模拟LLM输出。交付完整的Vibe Coding原型包。”
预期成果:简要说明+原型规格+提示包+任务看板+演示脚本;原型计划默认使用模拟数据及单一核心流程。
示例2(非工程师用户):“我是产品经理。使用
vibe-coding
帮我在45分钟内创建一个可点击的入职 checklist 应用原型。我需要一份给团队的演示脚本。”
预期成果:范围紧凑、使用模拟数据、按垂直切片开发、操作手册优化以保障演示可靠性。
边界示例:“通过Vibe Coding开发生产级支付后端并部署。”
回应:超出范围;建议仅做原型方案(模拟支付流程),明确所需的安全/工程负责人,并推荐单独的生产级开发计划。