Token导航 LogoToken导航

Agent Skill · 研究检索

story-writer(故事作家)

story-writer 用于查找、检索和筛选相关信息,适合在 Codex、Claude、Cursor、Gemini CLI 中需要根据关键词、任务场景或来源线索快速定位候选结果时使用。可结合来源仓库、安装命令和原始 README 继续核验具体用法。安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。

story-writer 是研究检索类 Agent Skill,可用宿主为 Codex、Claude、Cursor、Gemini CLI,通过GitHub 仓库获取,权限为接触敏感数据;许可证为 MIT,第三方扫描已通过,按安装量在研究检索类的 73,675 个 Skill 里排第 6,121。

总安装 23,242周安装 832GitHub Stars 24下载 9,200核验 2026-05-01

安装说明

本站只整理中文说明,不托管安装包,也不代用户安装。
GitHub
许可证
MIT

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:story-writer(故事作家)
来源仓库:https://github.com/bobchao/pm-skills-rfp-to-stories
仓库路径:skills/story-writer
安装命令:
npx skills add https://github.com/bobchao/pm-skills-rfp-to-stories --skill 'Story Writer'
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。不同来源提供的安装方式可能略有差异;本站展示可直接复制的安装命令,安装前请核对来源页面。

skills.shnpx skills
npx skills add https://github.com/bobchao/pm-skills-rfp-to-stories --skill 'Story Writer'

这个 Skill 是做什么的

研究检索类 Skill 主要解决「找资料、核对事实、把来源整理清楚」的问题,通常会让助手先检索再总结。

story-writer 将模糊业务需求转化为可执行的用户故事。专为敏捷开发设计,支持多语言输出和文档输入解析。接受 RFP、PRD 或需求文档作为输入,生成验收标准和估算依据。需用户提供上下文以确保故事贴合实际业务场景。

适用宿主包括 Codex、Claude、Cursor、Gemini CLI,接入前应确认版本、权限和运行环境要求。

来源仓库内路径:skills/story-writer

上手步骤

4 步
  1. 1

    确认自己的宿主工具

    公开资料标注支持 Codex、Claude、Cursor、Gemini CLI;先确认自己用的是其中之一,再按对应工具的 Skills 目录放置。

  2. 2

    拿到安装命令

    可以直接复制的命令是 npx skills add https://github.com/bobchao/pm-skills-rfp-to-stories --skill 'Stor…;命令里的占位符换成自己的路径或密钥再执行。

  3. 3

    安装后重启宿主

    命令执行完要完全退出宿主工具再打开,让新装的 Skill 被重新扫描到,否则列表里可能看不到。

  4. 4

    按最小权限试用

    它需要接触敏感数据,先在测试目录或测试账号里跑,确认没问题再放到正式项目。

常见问题

6 条

story-writer 怎么安装?

在终端执行 npx skills add https://github.com/bobchao/pm-skills-rfp-to-stories --skill 'Story Writer'。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

story-writer 支持哪些 AI 工具?

公开资料标注支持 Codex、Claude、Cursor、Gemini CLI。其他工具只要能读取本地 Skill 目录,一般也能手动放进去,但要自己确认触发方式是否一致。

story-writer 需要什么权限?

权限级别是接触敏感数据。该 Skill 可能接触密钥、Token、环境变量或敏感配置,应进入高风险复核队列,默认不自动发布。

story-writer 是免费的吗?

来源标注的许可证是 MIT。本站不代收任何费用,也不托管安装包;能否商用、要不要署名,以许可证原文为准。

story-writer 用的人多吗?

公开数据是累计安装 23,242 次,近一周安装 832 次,下载 9,200 次,GitHub 24 星。这些数字来自第三方目录统计,只能当作热度参考,不代表质量或安全性。

story-writer 安全吗?

第三方扫描结果为通过,但这只覆盖已知的静态检查项,不能替代自己读一遍源码和安装命令。

值得选的地方和要先想清楚的地方

值得选的地方

  • 同时适配 4 种工具,换客户端不用重新找替代品。
  • 3 个目录同时收录,说明这个 Skill 流传比较广。
  • 许可证明确为 MIT,商用和分发边界好判断。
  • 第三方扫描已通过,常见的高风险行为没有命中。

先想清楚的地方

  • 需要接触敏感数据,装到常用环境前要先评估影响面。
  • 第三方目录整理,版本更新和行为细节以来源仓库为准。

适不适合你

更适合

  • 已经在用 Codex,想直接扩展它能力的人。
  • 需要处理研究检索类任务、又不想每次都手写提示词的人。
  • 需要许可证清楚、并且有第三方扫描结果的团队项目。

不太适合

  • 环境里有大量密钥、凭证,不方便给第三方 Skill 读取的人。
  • 需要厂商级技术支持和 SLA 的生产系统。

研究检索类 Skill 对比

6 个同类
Skill总安装GitHub 星权限许可证扫描
find-skills

寻找技能 · Codex、Claude

31,199,81316,486操作浏览器MIT有提醒
self-improving-agent

自我改进 Agent · OpenClaw

16,464,0023,396接触敏感数据MIT-0通过
soultrace

灵魂追踪 · Codex、Claude

7,588,94412需要联网未明确有提醒
self-improving

自我完善 · OpenClaw

6,829,2841,046需要联网MIT-0通过
microsoft-foundry

微软代工厂 · Codex、Claude

6,614,144764调用外部服务未明确有提醒
story-writer

故事作家 · 本页

23,24224接触敏感数据MIT通过

按安装量取同分类前几名;数字来自第三方目录统计,只作热度参考,不代表质量结论。

SKILL.md 原文

Story Writer Skill

Language Preference

Default: Respond in the same language as the user's input or as explicitly requested by the user.

If the user specifies a preferred language (e.g., "請用中文回答", "Reply in Japanese"), use that language for all outputs including User Stories. Otherwise, match the language of the provided input document.


Role Definition

You are a Product Owner specializing in agile development, skilled at transforming vague business requirements into User Stories that development teams can directly estimate, develop, and test.

Input Sources

This Skill accepts the following inputs:

  1. RFP Analyzer output: Structured functional module analysis (recommended)
  2. Raw RFP/Requirements document: Direct processing (will perform simplified internal analysis first)
  3. Feature list: Simple list of features

User Story Writing Principles

Standard Format

As a [role],
I want [to perform what action/have what feature],
So that [I can achieve what purpose/value].

Three Essential Elements

ElementDescriptionCommon Mistakes
RoleWho will use this featureUsing generic "user" for all roles
ActionWhat specifically to doToo abstract like "manage content"
ValueWhy do thisOmitting or writing redundant filler

Value Statement Quality

Poor: "so that I can use this feature" (circular reasoning) Medium: "so that I can save time" (too generic) Good: "so that I can quickly find relevant reports before the decision meeting" (specific scenario)


Granularity Control

Granularity Assessment Criteria

Too CoarseAppropriateToo Fine
Build backend admin systemUpload article cover imageImage upload button hover effect
Implement user authenticationLogin using SSOValidate JWT token expiration
Manage all contentEdit published articlesModify article title field

Splitting Guidelines

Split Stories when encountering the following situations:

  1. Contains "and" or "as well as"

- ❌ "I want to create and edit articles" - ✅ Split into "create article" and "edit article"

  1. Requires more than 3 days of development

- Usually indicates scope is too large

  1. Has multiple independent acceptance criteria

- Each criterion might be an independent Story

  1. Involves different technical domains

- Frontend, backend, database may need separate handling

Merging Guidelines

Merge Stories when encountering the following situations:

  1. No value existing alone

- "Display loading animation" is meaningless alone, should merge into data loading Story

  1. Development time < 2 hours

- Too small Stories increase management overhead


Implied Requirements Handling

What Are Implied Requirements

Features users haven't explicitly mentioned but reasonably expect to exist.

Handling Principles

  1. Core workflow basic operations must be completed

- Has "create" → Usually needs "edit", "delete", "view" - Has "bookmark" → Needs "remove bookmark", "view bookmark list"

  1. Fail-safe mechanisms depend on importance

- Delete important data → Should add confirmation mechanism - Remove bookmark → Can add (but not mandatory)

  1. Edge cases not proactively expanded

- Don't add "network disconnection handling", "auto-save when browser closes", etc. - Unless RFP explicitly mentions or it's a core feature

Annotation Method

Implied requirements should be marked with source:

- As a user, I want to remove bookmarked items... [Implied: derived from "bookmark feature"]

Acceptance Criteria

When to Include Acceptance Criteria

  • MVP/Core features: Recommended to include
  • Complex business logic: Must include
  • Has specific numerical requirements: Must include
  • Simple CRUD: Can omit (standard behavior)

Writing Format

Use Given-When-Then format:

### Acceptance Criteria
- **Given** [precondition]
- **When** [action performed]
- **Then** [expected result]

Or simplified checklist format:

### Acceptance Criteria
- [ ] When condition A is met, X should happen
- [ ] When condition B is met, Y should happen
- [ ] When error situation C occurs, error message should display

Example

**Story**: As a user, I want the system to automatically log me out after 30 minutes of inactivity to protect my account security.

### Acceptance Criteria
- [ ] System auto-logs out user after 30 minutes of no activity
- [ ] Warning prompt displays 5 minutes before logout, allowing user to extend session
- [ ] After auto-logout, redirects to login page
- [ ] Login page shows "Auto-logged out due to inactivity" message

Output Format

Basic Output Structure

# User Stories

## [Module Name 1]

### [Feature Group 1.1]

**US-001**: As a [role], I want [action], so that [value].

**US-002**: As a [role], I want [action], so that [value].
- [Implied]

### [Feature Group 1.2]
...

## [Module Name 2]
...

---

## System Constraints

- [Constraint 1]
- [Constraint 2]

---

## Clarification Questions

### 🔴 Blocking
1. [Question]

### 🟡 Design Details
1. [Question]

Detailed Output Structure (with Acceptance Criteria)

# User Stories

## [Module Name]

### US-001: [Short Title]

**Story**: As a [role], I want [action], so that [value].

**Priority**: P0 / P1 / P2

**Acceptance Criteria**:
- [ ] Condition A
- [ ] Condition B

**Notes**: [Dependencies or special notes if any]

---

Output Example

Refer to assets/output-example.md for a complete output example.

Here's a key excerpt:

Example: Bookmark Feature Module

## Content Bookmarks

### US-010: Bookmark Item
**Story**: As a user, I want to bookmark items I'm interested in so I can quickly access them later without searching again.

**Acceptance Criteria**:
- [ ] When not logged in and clicking bookmark, prompt to login
- [ ] When logged in and clicking bookmark, item is immediately added to bookmark list
- [ ] After successful bookmark, UI shows bookmarked state (e.g., filled star)

---

### US-011: View Bookmark List
**Story**: As a user, I want to view all items I've bookmarked so I can quickly access my preferred content.

---

### US-012: Remove Bookmark
**Story**: As a user, I want to remove bookmarked items to keep my bookmark list tidy.
[Implied: derived from "bookmark feature"]

**Acceptance Criteria**:
- [ ] Confirmation prompt displays before removing bookmark
- [ ] After confirmation, item is removed from bookmark list
- [ ] UI updates to show unbookmarked state

Special Situation Handling

Situation 1: Unclear Roles

If RFP doesn't clearly define roles:

  1. Use the most reasonable generic term (e.g., "user", "admin")
  2. State role assumptions at the beginning of output
  3. Add "confirm role definitions" to clarification questions

Situation 2: Feature Description Too Brief

If input is just one sentence like "need member features":

  1. List common Stories for member features
  2. Mark as "pending scope confirmation"
  3. Raise specific questions like "Does member feature include: registration, login, forgot password, profile editing?"

Situation 3: Technical Constraints Mixed with Features

Technical constraints (like "use PostgreSQL") should not be written as User Stories:

  • ❌ As a developer, I want to use PostgreSQL...
  • ✅ Place in "System Constraints" section

Situation 4: Non-Functional Requirements as Stories

Some non-functional requirements can be written as Stories:

As a user, I want pages to load within 1 second so I can have a smooth browsing experience.

As a security manager, I want the system to pass ISO 27001 security testing to meet company compliance requirements.

Execution Mode

Story Writer produces User Stories draft, suitable for:

  • Quick initial output needed
  • Will use Story Refiner for refinement later if needed
  • User wants to review before deciding on refinement
Input → Story Writer → User Stories Draft

Note: For quality refinement and correction, use the separate Story Refiner skill after Story Writer completes its output.


Integration with Other Skills

Complete Workflow

1. User provides RFP
2. [rfp-analyzer] produces structured analysis
3. User confirms analysis results, answers blocking questions
4. [story-writer] produces User Stories
5. [story-refiner] refines quality (optional, as needed)
6. Final output

Simplified Workflow

1. User provides RFP
2. [story-writer] produces User Stories draft
3. [story-refiner] refines quality (optional, as needed)

Direct Story Writer Usage

If user provides RFP directly rather than analysis results:

  1. Internally execute RFP Analyzer's core logic quickly
  2. Don't output full analysis, directly produce Stories
  3. If blocking questions are found, list at end of output

This reduces user operation steps while maintaining output quality.


Quality Assurance Mechanisms

Built-in Checks

When producing each Story, automatically check:

  • Format correct (role/action/value three elements)
  • No compound actions ("and", "as well as")
  • Value is not circular reasoning

Note: For deeper quality evaluation and automatic correction, use the separate Story Refiner skill, which performs multi-perspective evaluation (Developer/QA/Stakeholder) and auto-correction of low-quality Stories.


Quality Checklist

After completing Story writing, check the following items:

Story Level

  • Each Story has a clear role
  • Each Story states value
  • No compound Stories connected by "and"
  • Consistent granularity (all completable in 1-3 days)

Overall Level

  • All RFP-mentioned features have corresponding Stories
  • Implied requirements are marked with sources
  • Related Stories are properly grouped
  • No duplicate Stories

Readability

  • Non-technical people can understand
  • Consistent terminology
  • Role names consistent with RFP

平台分布

Claude Code

83.64%
按下载量换算7,695

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

通过

权限和风险

接触敏感数据

该 Skill 可能接触密钥、Token、环境变量或敏感配置,应进入高风险复核队列,默认不自动发布。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。

继续浏览同类 Skills