New: 98,000+ skills indexed from GitHub repos with 100+ stars · every skill safety-checkedBrowse top skillsTop skills
Esc to closeBrowse all skills
Submit a skillBrowse skills
Submit a skillBrowse skills

Agent skill

Qiaomu AI PRD

by joeseesun in joeseesun/qiaomu-ai-prd

Turns a one-line product idea into an AI-implementable PRD with 11 chapters, a speed-read card, constraint layers and acceptance criteria, written Chinese-first.

MITAuto-check passedProduct & Project Management

Install Qiaomu AI PRD

skills CLI
$ npx skills add joeseesun/qiaomu-ai-prd --skill qiaomu-ai-prd -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI
$ gh skill install joeseesun/qiaomu-ai-prd qiaomu-ai-prd --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Claude Code skills documentation · loads skills from .claude/skills/

Facts

Skill name
qiaomu-ai-prd
GitHub stars
211
Token cost
~1.6k tokens
SKILL.md length
858 words
Files
10 (incl. scripts, references)
Skills in repo
1
Repo updated
First seen
Licence
MIT

At a glance

Turns a one-line product idea into an AI-implementable PRD with 11 chapters, a speed-read card, constraint layers and acceptance criteria, written Chinese-first.

  • Works in 12 steps: Parse the user input and optional mode… → Decide the likely product category,… → Decide the product's 硬约束, 推荐默认, and 发挥空间. → …
  • Turning a one-line app idea into a full PRD
  • SKILL.md covers Operating Mode, Workflow, Output Contract and Optional Modes, plus 2 more sections
  • Runs Python scripts from its folder; calls python3

What it does

This skill converts a vague product idea, even a single sentence, into a PRD that product managers, developers and AI coding assistants can act on. It behaves as a production-lite specification tool: it infers a conservative product direction rather than interrogating you, asks only when an answer would materially change category, platform, safety, legal risk, budget, data ownership or scope, and does not start implementation. Output is Chinese-first unless you request English.

The PRD follows a fixed contract of 11 chapters, preceded by a short speed-read card so an implementing agent can grasp the product quickly. Important instructions are labeled as hard constraints, recommended defaults or room for creativity. Each module gets ASCII UI and state diagrams, a normal flow, at least two failure paths and key product decisions, and the PRD adds a few overdelivery opportunities. Unverified competitor, API or package facts are marked unknown. A lint_prd.py script checks the result, and reference files cover modes and defaults, output quality and methodology.

When your agent uses it

  • Turning a one-line app idea into a full PRD
  • Writing developer handoff docs for an AI coding assistant
  • Scoping an MVP for a site, app or tool
  • Drafting acceptance criteria and metrics for a feature concept

Example prompts

  • “Write a PRD for an app that tracks houseplant watering.”
  • “把一个做读书笔记同步工具的想法写成 AI 可执行的 PRD”
  • “Draft an MVP plan and developer handoff doc for a recipe-sharing website.”

Requirements

  • Python, to run the lint_prd.py check

Workflow steps

12 steps, taken from the first numbered list in SKILL.md.

  1. Parse the user input and optional mode tags from references/modes-and-defaults.md.
  2. Decide the likely product category, target users, primary platform, and MVP surface.
  3. Decide the product's 硬约束, 推荐默认, and 发挥空间.
  4. Identify facts that must be verified, assumptions that can be used safely, and unknowns that must be represented honestly.
  5. Generate the PRD with the exact chapter contract in references/prd-methodology.md.
  6. For each module, include realistic ASCII UI/state diagrams, normal flow, at least two failure paths, states, dependencies, and 1-3 real…
  7. Add 超预期机会: 2-4 product moments that can make the implementation feel memorable without bloating P0.
  8. For differentiation and technical choices, explain structural causes and tradeoffs instead of saying competitors "did not think of it".
  9. Give numeric performance targets with measurement methods and degradation thresholds.
  10. Finish chapter 11 as a direct note to the implementing AI assistant using second person 你, including acceptance scripts it can run or…
  11. Run the self-check in references/output-quality.md before final output.
  12. If the PRD is saved to a file, run python3 scripts/lint_prd.py and fix any reported issue.

What it can do on your machine

Read from SKILL.md and the folder at commit dede555. It shows what the files ask for, not the result of running them.

  • Tool permissions

    Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

    From allowed-tools in the SKILL.md frontmatter.

  • Runs code

    Ships 1 file in scripts/ (Python), which the agent can run.

    Shell commands in SKILL.md call:

    • python3

    From the folder's file list and the shell code blocks in SKILL.md.

  • Network

    Links to these hosts (documentation or services it may open):

    • x.com
    • github.com

    From URLs in SKILL.md, links to its own repository left out.

  • Credentials

    Names no API keys, tokens, secrets or passwords.

    From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Qiaomu AI PRD loads about 1.6k tokens when it runs, and up to ~7.2k if it reads all its reference files. Until then it costs about 76 tokens; SKILL.md has 858 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~76
When it runs · the whole SKILL.md, loaded when a task matches
~1.6k
With references · SKILL.md plus every file in references/, read only if the agent opens them
~7.2k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check passed

The automated check found no risky patterns in SKILL.md.

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); the scripts in this folder are not scanned.

SKILL.md

The full file from joeseesun/qiaomu-ai-prd at commit dede555, republished under its MIT licence (© joeseesun). 858 words, ~1,641 tokens.

Download SKILL.mdSave it as .claude/skills/qiaomu-ai-prd/SKILL.md (or your agent's skills folder). This skill also uses 9 other files; get the full folder from GitHub.
name
qiaomu-ai-prd
description
Generate AI-implementable product requirement documents (PRDs) from one-line product ideas, vague app/site/tool requests, or feature concepts. Use when the user asks for PRD, 产品需求文档, 需求文档, AI 可执行 PRD, 产品规格, MVP 方案, app/site/tool planning, or developer handoff docs for AI coding assistants.

Qiaomu AI PRD

把一句模糊产品想法，写成产品经理、人类开发者和 AI 编程助手都能直接执行的 PRD。

Copyright (c) 向阳乔木 X: https://x.com/vista8 GitHub: https://github.com/joeseesun/

Operating Mode

Run as a production-lite product specification skill.

Default assumptions:

  • The user usually wants a finished PRD, not a questionnaire.
  • If the input is only one sentence, infer the best conservative product direction and continue.
  • Ask only when the answer would materially change product category, platform, safety, legal risk, budget, data ownership, or implementation scope.
  • When a choice is needed, make the best default decision and give the reason inside the relevant chapter.
  • Do not start implementation unless the user explicitly asks; this skill produces the PRD.
  • Write Chinese-first unless the user asks for English.
  • Output all 11 required chapters in order. Do not skip chapters, even in compact mode.
  • Add an AI 速读卡 before the chapters so an implementing agent can grasp the product in 10 lines or fewer.
  • Keep implementation details out unless they affect product behavior, architecture risk, data contracts, verification, or AI handoff.
  • Do not invent current competitor, API, platform, or package facts. If current facts matter and cannot be verified, mark them as unresolved or use 未知.
  • Treat fuzzy product words as direction, not proof. Translate them into concrete UI states, measurable targets, outputs, and acceptance criteria.
  • Separate each important instruction into 硬约束, 推荐默认, or 发挥空间 so implementation agents know what must hold and where they can improve freely.

Workflow

  1. Parse the user input and optional mode tags from references/modes-and-defaults.md.
  2. Decide the likely product category, target users, primary platform, and MVP surface.
  3. Decide the product's 硬约束, 推荐默认, and 发挥空间.
  4. Identify facts that must be verified, assumptions that can be used safely, and unknowns that must be represented honestly.
  5. Generate the PRD with the exact chapter contract in references/prd-methodology.md.
  6. For each module, include realistic ASCII UI/state diagrams, normal flow, at least two failure paths, states, dependencies, and 1-3 real product decisions or 无.
  7. Add 超预期机会: 2-4 product moments that can make the implementation feel memorable without bloating P0.
  8. For differentiation and technical choices, explain structural causes and tradeoffs instead of saying competitors "did not think of it".
  9. Give numeric performance targets with measurement methods and degradation thresholds.
  10. Finish chapter 11 as a direct note to the implementing AI assistant using second person 你, including acceptance scripts it can run or manually verify.
  11. Run the self-check in references/output-quality.md before final output.
  12. If the PRD is saved to a file, run python3 scripts/lint_prd.py <file> and fix any reported issue.

Output Contract

When the user gives a product idea, output the PRD directly. Use this order:

  1. # [产品名] PRD
  2. ## AI 速读卡
  3. ## 第一章：产品概述
  4. ## 第二章：整体布局与导航
  5. ## 第三章：核心模块详细设计
  6. ## 第四章：超越竞品的差异化功能
  7. ## 第五章：数据模型
  8. ## 第六章：技术架构
  9. ## 第七章：交互细节
  10. ## 第八章：导出与输出系统
  11. ## 第九章：开发优先级
  12. ## 第十章：性能指标
  13. ## 第十一章：开发者交接说明

Do not add a long preface. If assumptions are needed, place them inside the relevant chapter, usually 1.3 可行性边界, module 待决问题, or 第十一章 d) 已知的未知项.

Show full SKILL.md (371 more words)Show less

Optional Modes

Recognize these tags anywhere in the user request:

  • [深度模式]: add boundary-case analysis to each major module.
  • [精简模式]: keep every chapter, but focus detailed design on P0; mark lower tiers as 待扩展.
  • [前端视角]: add component decomposition and state-management guidance where product-relevant.
  • [后端视角]: add API design and database schema where product-relevant.
  • [移动优先]: make all layout diagrams mobile-first unless the product is clearly desktop-only.
  • [竞品深挖]: deepen competitor weakness analysis and product blind-spot reasoning.
  • [商业化]: add pricing, paid feature, and monetization implications where appropriate.
  • [开源友好]: prefer permissive open-source libraries, especially MIT, when the choice does not harm the product.

See references/modes-and-defaults.md for how to combine modes.

Quality Bar

A strong PRD from this skill:

  • makes product decisions instead of pushing every ambiguity to the user
  • gives an implementing agent a short AI 速读卡
  • distinguishes 硬约束, 推荐默认, and 发挥空间
  • contains realistic ASCII diagrams with actual labels and representative content
  • names meaningful competitor differences instead of filling a comparison table with obvious parity
  • defines module states, data flows, failure paths, and open decisions
  • includes a small set of 超预期机会 that invite tasteful implementation beyond the baseline
  • uses data structures with commented JSON fields and a top-level version
  • explains technical choices and package-size uncertainty honestly
  • explains when a technical choice is replaceable and what must remain invariant
  • prioritizes by user behavior impact, not implementation difficulty
  • turns performance expectations into exact numbers and measurement methods
  • tells the implementing AI what to build first, what not to reinterpret, what to freely improve, what remains unknown, and how to verify the first build

Reject or revise a PRD that:

  • leaves placeholders such as [产品名], 按钮 A, TODO, or 待补充
  • uses vague performance language such as 快, 流畅, 轻量, or 可扩展 instead of numbers
  • claims impossible browser, iOS, Android, web, AI model, or export capabilities
  • invents competitor facts, package sizes, or platform limits
  • has no honest known-unknown item in chapter 11
  • lacks 验收剧本 for implementation verification
  • lists P0 as a wishlist instead of the smallest usable product

Reference Files

  • references/prd-methodology.md: the required 11-chapter PRD structure and detailed generation rules.
  • references/modes-and-defaults.md: lazy-user defaults, optional modes, question policy, and uncertainty handling.
  • references/output-quality.md: output self-check and common failure patterns.
  • scripts/lint_prd.py: lightweight checker for required chapters, unresolved placeholders, vague performance terms, and structural omissions.

© joeseesun, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

SKILL.md and 9 other files (scripts, references) in the repository root of joeseesun/qiaomu-ai-prd.

  • SKILL.md
  • .gitignore
  • LICENSE
  • README.md
  • agents/interface.yaml
  • manifest.json
  • references/modes-and-defaults.md
  • references/output-quality.md
  • references/prd-methodology.md
  • scripts/lint_prd.py

Open the folder on GitHubat commit dede555

Compare with similar skills

Qiaomu AI PRD next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

Qiaomu AI PRD compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Qiaomu AI PRD this skilljoeseesun/qiaomu-ai-prd211—~1.6kAutomated safety check: PassMIT
Ouroboros PM InterviewQ00/ouroboros6.2k—~5.7kAutomated safety check: PassMIT
MVP Product Requirements WriterKhazP/vibe-coding-prompt-template3.1k—~455Automated safety check: PassMIT
User Alignment and Agent-Ready PRDstryproduck/produck-skills511—~5.3kAutomated safety check: PassApache-2.0
Code to PRDalirezarezvani/claude-skills28k1 repos~4.9kAutomated safety check: PassMIT
Requirements From HTML Mocksimbue-ai/bouncer399—~410Automated safety check: PassAGPL-3.0

Similar skills

  • Ouroboros PM Interview

    Q00/ouroboros

    Runs a guided product-manager interview that classifies each question automatically and produces a Product Requirements Document.

    6.2k GitHub stars~5.7k tokensUpdated 4 days ago
    Product & Project ManagementAuto-check passed

  • MVP Product Requirements Writer

    KhazP/vibe-coding-prompt-template

    Writes or revises an MVP product requirements document, sized to the product, with observable acceptance criteria and no repeated questions.

    3.1k GitHub stars~455 tokensUpdated 6 days ago
    Product & Project ManagementAuto-check passed

  • User Alignment and Agent-Ready PRDs

    tryproduck/produck-skills

    Turns a vague feature request into a written spec with scope, phases, acceptance criteria and do-not-do limits that a coding agent can follow without guessing.

    511 GitHub stars~5.3k tokensUpdated 1 mo ago
    Product & Project ManagementAuto-check passed

  • Code to PRD

    alirezarezvani/claude-skills

    Reverse-engineers a frontend, backend or fullstack codebase into a product requirements document with per-page docs, an enum dictionary and an API inventory.

    28k GitHub starsUsed in 1 repo~4.9k tokens
    Product & Project ManagementAuto-check passed

  • Requirements From HTML Mocks

    imbue-ai/bouncer

    Turns an HTML mock iteration session into a requirements document that describes only user-facing behavior.

    399 GitHub stars~410 tokensUpdated 3 days ago
    Product & Project ManagementAuto-check passed

  • Creating Issues

    opsmill/infrahub

    Turns a single feature idea, improvement, or bug into ONE well-structured GitHub issue.

    534 GitHub stars~1.2k tokensUpdated yesterday
    Product & Project ManagementAuto-check passed

Categories

Questions about Qiaomu AI PRD

What does Qiaomu AI PRD do?

Turns a one-line product idea into an AI-implementable PRD with 11 chapters, a speed-read card, constraint layers and acceptance criteria, written Chinese-first. This skill converts a vague product idea, even a single sentence, into a PRD that product managers, developers and AI coding assistants can act on. It behaves as a production-lite specification tool: it infers a conservative product direction rather than interrogating you, asks only when an answer would materially change category, platform, safety, legal risk, budget, data ownership or scope, and does not start implementation.

When should I use Qiaomu AI PRD?

Qiaomu AI PRD fits situations like: turning a one-line app idea into a full PRD; writing developer handoff docs for an AI coding assistant; scoping an MVP for a site, app or tool; drafting acceptance criteria and metrics for a feature concept.

How do I install Qiaomu AI PRD in Claude Code?

Run `npx skills add joeseesun/qiaomu-ai-prd --skill qiaomu-ai-prd -a claude-code`. Or copy the skill folder (the joeseesun/qiaomu-ai-prd repository) into .claude/skills/qiaomu-ai-prd in your project. Claude Code loads it when a task matches its description.

How do I install Qiaomu AI PRD in Codex?

Run `npx skills add joeseesun/qiaomu-ai-prd --skill qiaomu-ai-prd -a codex`. Or copy the skill folder (the joeseesun/qiaomu-ai-prd repository) into .agents/skills/qiaomu-ai-prd in your project. Codex loads it when a task matches its description.

Can I use Qiaomu AI PRD in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add joeseesun/qiaomu-ai-prd --skill qiaomu-ai-prd -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/qiaomu-ai-prd, .gemini/skills/qiaomu-ai-prd, .github/skills/qiaomu-ai-prd and .opencode/skills/qiaomu-ai-prd in your project.

What does Qiaomu AI PRD need to run?

Going by SKILL.md and its folder, Qiaomu AI PRD needs Python for the scripts in its folder and the command-line tools its instructions call (python3). Our summary lists: Python, to run the lint_prd.py check.

Does Qiaomu AI PRD access the network?

SKILL.md names 2 domains. As links in the text: x.com and github.com. This is read from the text; nothing was executed.

Is Qiaomu AI PRD safe to install?

Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. The check reads SKILL.md only: the scripts in the folder are not scanned, so read them before running anything.

What licence does Qiaomu AI PRD use?

Qiaomu AI PRD is published under the MIT licence (from the LICENSE file in the skill folder). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Qiaomu AI PRD use?

About 1.6k tokens (SKILL.md is roughly 6.6k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full. Its references folder adds about 5.6k tokens, read only when the agent opens those files.

What are the alternatives to Qiaomu AI PRD?

Skills that share tags, products or a category with Qiaomu AI PRD: Ouroboros PM Interview (Q00/ouroboros, 6.2k stars), MVP Product Requirements Writer (KhazP/vibe-coding-prompt-template, 3.1k stars), User Alignment and Agent-Ready PRDs (tryproduck/produck-skills, 511 stars) and Code to PRD (alirezarezvani/claude-skills, 28k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Qiaomu AI PRD?

joeseesun (a GitHub user) maintains it in joeseesun/qiaomu-ai-prd, which has 211 GitHub stars. The repository was last updated on June 12, 2026.

Source: joeseesun/qiaomu-ai-prd on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.