product-requirement-delivery

SKILL商业运营
商业运营 未标注权限 许可证不明确
安装这个技能

Convert a lightweight product request plus related signed-in pages into a product-confirmed requirements baseline published as one Feishu document. Use when a product manager asks Codex to investigate existing pages, clarify roles and business rules, capture complete route-aware screenshots, define user stories and exception scenarios, obtain product confirmation, and write the final requirement to Feishu for downstream development and testing. Also use for requests such as “按现有页面出需求”“把需求写到飞书”“补齐角色、用户故事、异常场景和验收标准”.

一句话选型结论

按资料字段整理

product-requirement-delivery是商业运营类的 Agent Skill,安装方式未标注,权限级别未标注,许可证不明确,有可直接复制的安装命令。技能本质是给模型读的说明书,效果取决于模型与客户端;装完先用一个小任务验证,再放进正式流程。

分类商业运营 安装方式未标注 权限级别未标注权限 许可证许可证不明确
累计安装 未标注 本周 未标注
GitHub 星标 未标注 仓库 未标注
下载量 未标注 技能市场下载数
权限级别 未标注权限 安装前请确认风险范围
许可证 许可证不明确 未说明

安装

未标注
在支持 Skills 的客户端里执行,装完问一句「你有哪些技能」验证
skillhub install user_926b9ad6/product-requirement-delivery-1
  1. 1
    确认客户端支持

    确认你的客户端支持 Skills(Claude、Codex、Cursor 等),并升级到较新的版本。

  2. 2
    拿到技能文件

    执行上面的命令安装;也可以把仓库里的 SKILL.md 手动放进客户端的 skills 目录。

  3. 3
    刷新技能列表

    重新打开客户端,或刷新技能列表,让新放进来的技能被加载。

  4. 4
    验证是否生效

    问一句「你有哪些技能」,确认列表里出现它,再拿一个小任务试跑。

技能档案

商业运营
字段来自技能仓库与公开索引,中文说明由本站标注
作者 / 维护方user_926b9ad6仓库 未标注 · 版本 未标注
宿主客户端未标注技能可被这些客户端加载
安装方式未标注安装包 community
权限级别未标注权限安装前请确认风险范围
许可证未说明许可证不明确,商用前请自行确认条款
仓库可用性available技能仓库当前是否可访问
最近核验未标注上游最后核验时间 · 数据更新 1791675860
正文体量7.3 KBSKILL.md 原文大小,越大说明说明书写得越细

典型使用场景

3 类
按分类整理的常见用法,具体以 SKILL.md 为准
写运营方案

按目标产出活动方案与执行清单。

分析数据与指标

从报表里找问题、给结论。

整理客户与销售材料

把客户信息与话术整理成可直接用的材料。

选择建议

装之前先看这几条
按条目字段给出的判断,不替你决定用不用

值得选的地方

  • 有可直接复制的安装命令,不用自己翻文档找入口。

先想清楚的地方

  • 没有标注权限级别,实际能做什么要装好后确认。
  • 许可证不明确,商用或二次分发前先到仓库确认条款。
  • 没有做站内核验,内容以作者仓库为准。
  • 技能只是说明书,实际效果取决于模型与客户端,装完先用小任务试一遍。

更适合

  • 需要在商业运营场景里给模型补一套做法的人。
  • 愿意先读一遍 SKILL.md、确认它会让模型做什么的人。

不太适合

  • 对许可证要求很严、必须明确可商用的人。
  • 只想装完就自动变强、不打算看说明书的人。

同类对比

6 个同类
同分类里挑最接近的几条,字段以各自详情页为准
技能权限级别安装方式累计安装宿主客户端安装命令
product-requirement-delivery当前 未标注权限 未标注 未标注 未标注 有
MX 股票筛选器 联网访问 ClawHub 市场 79.9 万 OpenClaw 有
交易 未标注权限 ClawHub 市场 58.1 万 OpenClaw 有
营销技巧 只读 ClawHub 市场 54.0 万 OpenClaw 有
百灵鸟好 联网访问 GitHub 仓库 52.6 万 Codex、Claude、Cursor 有
股票信息浏览器 接触敏感数据 ClawHub 市场 40.5 万 OpenClaw 有

SKILL.md 全文

7.3 KB
模型实际读到的说明书原文,来自技能仓库

Product Requirement Delivery

Turn a lightweight requirement card into one clear, reviewable Feishu requirement document. Keep the product manager responsible for business decisions. Do not require the product manager to prescribe code, APIs, database design, development plans, or test plans.

Required workflow

1. Audit the intake

Read references/intake-card.md. Extract the requirement name, related pages, current problem, target users, desired result, known rules, exclusions, and references.

Do not ask the user to write a full PRD. Discover page facts independently. Propose low-risk business defaults. Ask only questions whose answers materially change roles, permissions, visible behavior, state changes, data meaning, exception handling, or acceptance results.

2. Investigate the real product

Open every related page and inspect its navigation, fields, filters, states, operations, terminology, permissions visible to the current account, empty states, and adjacent flows. Use the user's signed-in Chrome session when login state matters, after reading the applicable Chrome/browser skill.

Read and obey references/ui-evidence-rules.md. For every page, capture both route evidence and complete page evidence. Preserve an unannotated original; annotations are supplemental.

Classify statements as:

  • Observed: directly verified on the live page.
  • User stated: supplied by the user.
  • Recommended: proposed for product confirmation.
  • Unverified: inaccessible or not provable from the current page state.

Never fabricate a replacement interface. If a state is unavailable, define the required business behavior and mark placement or current behavior unverified.

3. Draft the requirements baseline

Create <requirement-name>-需求基线-v0.1.md using references/fact-source-schema.md. The local Markdown is the auditable source used to publish the Feishu document.

The draft must include:

  • roles and permission boundaries;
  • core user stories and preconditions;
  • explicit normal-flow rules;
  • state, default, editability, visibility, effective-time, and historical-data rules;
  • exception and boundary scenarios;
  • scope in/out;
  • observable Given/When/Then acceptance criteria;
  • complete URLs and screenshot evidence.

Use stable IDs: ROLE-*, US-*, REQ-*, EX-*, AC-*, and UV-*. Every AC must reference at least one REQ, US, or EX.

Run:

python scripts/validate_fact_source.py <requirement.md> --expected-status draft

4. Audit ambiguity before review

For every operation, make the following answerable without guessing:

  1. Which role can see or perform it?
  2. What preconditions and business states allow it?
  3. What is the default value and which values are legal?
  4. What exactly triggers success, and when does the result become visible?
  5. What changes, what remains unchanged, and how is history treated?
  6. What happens for no permission, invalid state, empty/not-found data, duplicate action, stale or concurrent modification, timeout, partial batch failure, and retry?
  7. What feedback does the user receive on success and failure?
  8. Which behavior is explicitly out of scope?

Do not use vague expressions such as “按需”“适当”“视情况”“正常处理” unless they are immediately converted into objective conditions and results.

5. Stop for product confirmation

Present only unresolved business decisions, each with a recommended default and impact. Do not publish the final Feishu document until the product manager explicitly confirms the rules.

6. Freeze the baseline

After confirmation:

  • change status to 已确认(可发布到飞书);
  • set version to v1.0 or later;
  • record confirmation date and decisions;
  • remove tentative language from confirmed rules;
  • keep inaccessible page facts as Unverified without weakening confirmed business behavior.

Run the validator with --expected-status confirmed.

7. Publish one Feishu requirement document

Read references/deliverable-contract.md. Then read and follow the lark-doc skill and its required shared authentication instructions.

  • If the user provides an existing Feishu document, update that document.
  • Otherwise create one document titled <requirement-name>-产品需求-v<version> in the user-specified folder, or the authenticated user's default document location when no folder is specified.
  • Publish the confirmed source as readable Feishu content, not as an attached DOCX.
  • Upload and insert the original complete screenshots and, when useful, annotated copies.
  • Put the copyable full URL next to every screenshot group even when the browser address bar is visible.
  • Do not create development-Agent packages, test-Agent packages, interface specifications, database designs, or implementation checklists unless the user separately asks for them.

8. Read back and verify

Fetch the published Feishu document and verify:

  • title and version;
  • all required headings and IDs;
  • roles, user stories, rules, exceptions, scope, and ACs;
  • every associated page URL;
  • every required screenshot is embedded and readable;
  • no confirmed rule reverted to tentative wording.

Create an audit manifest from references/audit-manifest.example.json, save the fetched Markdown, and run:

python scripts/audit_delivery.py --manifest <manifest.json> --source <confirmed.md> --published <fetched.md>

Fix the Feishu document and repeat the read-back audit until it passes.

Non-negotiable gates

  • One confirmed requirements baseline governs the Feishu document.
  • No final Feishu publication before product confirmation.
  • Every involved role has an explicit permission boundary.
  • Every core business goal has at least one user story and one acceptance criterion.
  • Normal, state, permission, empty/not-found, duplicate/retry, concurrency/stale-data, failure, and batch-partial-failure behavior are either defined or explicitly marked not applicable with a reason.
  • Route evidence includes the browser address bar; page evidence covers the complete relevant page with continuous context.
  • Screenshots never replace copyable full URLs.
  • No fabricated product UI or unverified technical implementation presented as fact.
  • No mandatory development or test package is produced from this skill.
  • Preserve the user's existing files and unrelated workspace changes.

Handoff

Lead with the Feishu document link and confirmed version. State any remaining Unverified page facts and what evidence was unavailable. Do not expose internal work files unless requested.

常见问题

5 问
答案由本站依据条目字段整理
product-requirement-delivery怎么安装?

在支持 Skills 的客户端里执行:skillhub install user_926b9ad6/product-requirement-delivery-1。安装方式为未标注。装完刷新技能列表或重开客户端。

product-requirement-delivery支持哪些客户端?

资料没有标注宿主客户端,但只要客户端支持 Skills(Claude、Codex、Cursor 等),一般都能加载。

product-requirement-delivery权限高吗,安全吗?

权限级别为未标注权限。安装前先看它会读取或写入哪些内容。

product-requirement-delivery收费吗?

本站不代收费用,也没有标注具体价格。是否收费取决于技能本身和所用的模型或客户端,正式使用前按作者说明确认。

product-requirement-delivery用什么许可证?

许可证标注为 未说明(许可证不明确)。商用或二次分发前请到仓库确认条款原文。

热门专题

按场景挑技能
全部专题 →

技能内容以作者仓库为准;分类、权限级别与许可证状态的中文说明为中文整理,最近核验 未标注。