Setup
On first use, read setup.md to align activation behavior, current storytelling goal, and audience context without delaying the immediate task.
When to Use
User needs to explain, persuade, or teach through narrative and wants a story that is coherent, specific, and emotionally engaging.
Use this skill for product stories, founder narratives, case studies, speeches, long-form writing, and short content adaptations.
Architecture
Memory lives in ~/storytelling/. See memory-template.md for baseline structure.
CODEBLOCK0
Quick Reference
Load only the smallest file needed to solve the current bottleneck.
| Topic | File |
|---|
| Setup and activation behavior | INLINECODE3 |
| Memory structure and status model |
memory-template.md |
| Arc design and narrative sequencing |
story-arc-map.md |
| Scene construction and detail layering |
scene-design.md |
| Format-specific adaptations and compression |
rewrite-modes.md |
| Voice calibration and consistency checks |
voice-consistency.md |
Data Storage
Local notes in ~/storytelling/ may include:
- - audience assumptions, emotional target, and success criteria
- bank of anecdotes, examples, and proof artifacts
- draft variants, edit decisions, and rejected directions
- reusable hooks, openings, transitions, and closings
Core Rules
1. Anchor Every Story to One Audience Outcome
Define one concrete outcome before drafting:
- - understand a complex idea
- believe a claim
- make a decision
- remember a key message
If the desired audience shift is not explicit, the story drifts into pleasant but ineffective prose.
2. Build a Causal Arc, Not a Topic List
Force each section to answer one of these transitions:
- - context -> tension
- tension -> decision
- decision -> action
- action -> result
If two consecutive sections do not have a causal bridge, add one or remove one section.
3. Use Specific Evidence at the Point of Highest Skepticism
Place proof where disbelief is most likely:
- - before major claims
- after bold promises
- inside turning points
Evidence can be data, concrete examples, constraints, trade-offs, or observed outcomes. Generic claims without proof collapse trust.
4. Control Emotional Pace with Scene Density
Alternate tight scenes and high-level summaries:
- - tight scene for empathy and credibility
- summary passage for speed and direction
Overusing scenes slows momentum. Overusing summaries removes emotional impact.
5. Separate Drafting from Judgment
Run two explicit passes:
- - pass A: generate material without heavy self-editing
- pass B: cut, reorder, and sharpen for clarity
Mixing ideation and critique in one pass usually creates safe, flat narratives.
6. Adapt Format Without Losing Core Story Logic
For every channel version, preserve:
- - core conflict
- key decision
- core proof
- final implication
Short formats require compression, not simplification into vague slogans.
7. End with a Single Clear Action or Reflection
Close with one explicit endpoint:
- - what to do next
- what to believe now
- what to watch for
A strong ending turns narrative quality into practical impact.
Storytelling Traps
- - Opening with background before stakes -> audience attention drops before the story starts.
- Adding too many subplots -> the central message becomes untraceable.
- Using abstract adjectives instead of concrete moments -> no mental image, low recall.
- Explaining every detail chronologically -> pacing slows and key turning points disappear.
- Rewriting tone without rechecking logic -> polished text with broken argument flow.
- Ending with broad inspiration only -> no decision, no behavior change, no result.
Security & Privacy
Data that leaves your machine:
- - none by default from this instruction set
- only user-approved outbound requests when the user explicitly asks for external research
Data that stays local:
- - storytelling context and iterative notes under INLINECODE10
- draft structure choices and narrative experiments
This skill does NOT:
- - request secrets, passwords, or private credentials
- make hidden network calls
- perform irreversible actions automatically
Scope
This skill ONLY:
- - structures narrative strategy for clarity, persuasion, and memorability
- improves story logic, evidence placement, pacing, and adaptation
- supports iterative drafting with explicit quality checks
This skill NEVER:
- - fabricate facts or testimonials
- claim outcomes that evidence cannot support
- replace domain review when factual or legal accuracy is required
Related Skills
Install with
clawhub install <slug> if user confirms:
- -
writing - Improve writing quality with clearer structure and stronger revision passes. - INLINECODE13 - Connect stories to audience segments, funnel stages, and distribution plans.
- INLINECODE14 - Create consistent narrative components for UI and product communication flows.
- INLINECODE15 - Build context-rich historical narratives with chronology and source-aware framing.
- INLINECODE16 - Turn transcript material into tighter narrative scripts and summaries.
Feedback
- - If useful: INLINECODE17
- Stay updated: INLINECODE18
初始设置
首次使用时,请阅读 setup.md 以对齐激活行为、当前叙事目标及受众背景,无需延迟当前任务。
使用时机
当用户需要通过叙事来解释、说服或教导,并希望故事具备连贯性、具体性和情感感染力时使用。
适用于产品故事、创始人叙事、案例研究、演讲、长文写作及短内容改编。
架构
记忆文件存放于 ~/storytelling/ 目录。基础结构请参见 memory-template.md。
text
~/storytelling/
|-- memory.md # 受众画像、活跃叙事目标及约束条件
|-- story-bank.md # 可复用的故事、场景及论据要点
|-- messaging-pillars.md # 核心主题、承诺及支撑证据
-- edit-log.md # 草稿迭代、决策过程及变更记录
快速参考
仅加载解决当前瓶颈所需的最小文件。
| 主题 | 文件 |
|---|
| 初始设置与激活行为 | setup.md |
| 记忆结构与状态模型 |
memory-template.md |
| 故事弧设计与叙事序列 | story-arc-map.md |
| 场景构建与细节层次 | scene-design.md |
| 格式适配与压缩 | rewrite-modes.md |
| 语调校准与一致性检查 | voice-consistency.md |
数据存储
~/storytelling/ 中的本地笔记可包含:
- - 受众假设、情感目标及成功标准
- 轶事、案例及证据素材库
- 草稿变体、编辑决策及被否决的方向
- 可复用的钩子、开场白、过渡句及结尾
核心规则
1. 每个故事锚定一个受众结果
在起草前明确一个具体结果:
如果期望的受众转变不明确,故事将沦为悦耳但无效的散文。
2. 构建因果弧,而非主题列表
强制每个部分回答以下过渡之一:
- - 背景 -> 张力
- 张力 -> 决策
- 决策 -> 行动
- 行动 -> 结果
若连续两部分缺乏因果桥梁,则添加或移除其中一个部分。
3. 在最高怀疑点使用具体证据
将证据置于最可能产生怀疑之处:
证据可以是数据、具体案例、约束条件、权衡取舍或观察结果。缺乏证据的泛泛之谈会摧毁信任。
4. 通过场景密度控制情感节奏
交替使用紧凑场景与高层级总结:
- - 紧凑场景用于共情与可信度
- 总结段落用于速度与方向
过度使用场景会拖慢节奏。过度使用总结会削弱情感冲击。
5. 将起草与评判分离
执行两个明确的轮次:
- - 第一轮:生成素材,不做过多自我编辑
- 第二轮:删减、重排、精炼以提升清晰度
将构思与批评混为一谈,通常会产生安全但平淡的叙事。
6. 适配格式而不丢失核心故事逻辑
针对每个渠道版本,保留:
短格式需要压缩,而非简化为模糊的口号。
7. 以单一明确的行动或反思结尾
以一个明确的终点收尾:
强有力的结尾能将叙事质量转化为实际影响力。
叙事陷阱
- - 在揭示利害关系前先介绍背景 -> 故事尚未开始,观众注意力已流失。
- 添加过多支线情节 -> 核心信息变得难以追踪。
- 使用抽象形容词而非具体时刻 -> 缺乏画面感,记忆度低。
- 按时间顺序解释每个细节 -> 节奏拖沓,关键转折点被淹没。
- 重写语调而不重新检查逻辑 -> 文字精炼但论证链条断裂。
- 仅以宽泛的激励结尾 -> 无决策、无行为改变、无结果。
安全与隐私
离开本机的数据:
- - 默认情况下,本指令集不产生任何数据外传
- 仅当用户明确要求外部研究时,才发送经用户批准的出站请求
保留在本地数据:
- - ~/storytelling/ 目录下的叙事上下文与迭代笔记
- 草稿结构选择与叙事实验
本技能不会:
- - 请求机密信息、密码或私人凭证
- 发起隐藏的网络调用
- 自动执行不可逆操作
范围
本技能仅:
- - 为清晰度、说服力与记忆度构建叙事策略
- 优化故事逻辑、证据布局、节奏与适配
- 通过明确的质量检查支持迭代起草
本技能绝不:
- - 捏造事实或推荐语
- 声称证据无法支持的结果
- 在需要事实或法律准确性时替代领域审查
相关技能
在用户确认后,使用 clawhub install 安装:
- - writing - 通过更清晰的结构和更强的修订轮次提升写作质量。
- content-marketing - 将故事与受众细分、漏斗阶段及分发计划相连接。
- storybook - 为 UI 和产品沟通流程创建一致的叙事组件。
- history - 构建富含背景的历史叙事,兼顾时间线与来源意识。
- youtube-video-transcript - 将转录素材转化为更紧凑的叙事脚本与摘要。
反馈
- - 如有帮助:clawhub star storytelling
- 获取更新:clawhub sync