Storytelling Skill
lazyfoxjumps/storytelling-skill
Transforma jargão técnico em prosa narrativa humanizada e clara
Ver no GitHubUma skill para Claude que transforma conteúdo técnico, repleto de jargão ou estruturado em listas em prosa narrativa contínua e humanizada. Preserva o significado original enquanto torna o texto legível como se escrito por uma pessoa que realmente entende o assunto, em vez de slides de consultant.
Instalação
npx convez-ai lazyfoxjumps/storytelling-skill1 comando no terminal: detecta o formato do repositório (marketplace ou plugin avulso) e instala via Claude Code. Requer Node 18+ e o Claude Code instalado.
Métricas
Casos de uso
- Simplificar documentação técnica para públicos leigos
- Reescrever relatórios e dados em formato narrativo engajante
- Traduzir conteúdo acadêmico ou especializado em linguagem acessível
Veredito

Skill bem estruturada e pragmática que resolve um problema real: tornar conteúdo denso acessível sem perder precisão. A ênfase em fidelidade ao original e tom honesto a diferencia de ferramentas genéricas de simplificação. Útil para qualquer profissional que precisa comunicar ideias complexas com clareza.
README
A tradução para o português deste README ainda está na fila.
Storytelling
A Claude Code skill that rewrites complex, technical, or structured content into continuous narrative prose. The reader should feel like they are being written to by a thoughtful person who actually understands the material, not briefed by a consultant presenting slides.
Paragraphs over bullet points. Sentences over headers. A human voice over a formal one.
What it does
Drop in dense, jargon-heavy, or list-shaped content and get back a unified piece of prose that preserves the original meaning but reads like something a real person would want to read. Works across any domain: astrology, finance, medicine, law, software, science, or anything that benefits from being translated out of expert-speak.
When it triggers
The skill activates when you ask Claude to:
- "Tell the story" of something
- Rewrite a technical document in plain language
- Produce a reading or report in "letter style" or "storytelling format"
- "Make this sound human", "write this like you're talking to me", "turn this into a story"
- "Explain this like I'm five", "simplify this", or "make this easier to understand"
- Or when you paste dense content and ask for it to be rephrased
Even when only one section of a longer piece needs simplifying, the skill rewrites the whole thing in a unified voice so the tone does not lurch.
Core behaviors
Plain language. Every piece of jargon gets translated into everyday words. No assumed acronyms, no industry shorthand. When a concept needs a label, it is introduced naturally in context.
Real analogies. Abstract ideas get illustrated with comparisons drawn from everyday human experience: relationships, routines, seasons, decisions, money, time. The kind of analogy you would actually use in conversation, not a textbook example.
Strict fidelity. The original meaning and structure of the source are preserved exactly. Claims are not reordered, conclusions are not reframed, and new ideas are not introduced. If something in the source is unclear, it is surfaced as a question rather than guessed at.
Prose-first formatting. Bullet points and lists get folded into sentences and paragraphs. Headers are used sparingly, only where they genuinely help a reader navigate. No em-dashes or en-dashes are used as punctuation; commas, colons, parentheses, or new sentences replace them.
Length discipline. Output length matches input complexity. Short input, short output. Long, dense input earns a longer response. No padding, no truncation.
Honest tone. Warm and direct, written to "you". Dry humor is welcome when it fits naturally. If the source makes a weak or self-serving claim, the skill names it plainly rather than softening it into agreement.
Input handling
You can pass in any of:
- Pasted text or a file path to rewrite
- A topic or concept to explain from scratch in this voice
- A source document plus additional context or a question
Before writing, the skill asks one calibration question: "Who is the intended reader?" The answer sets the register: a close friend needs less scaffolding than a stranger, and a client reads differently than a curious teenager. If you tell it to skip the question and just write, it will.
Example
Before (technical):
Mercury in Gemini trine Neptune in Aquarius creates a 120-degree harmonious aspect between the natal chart ruler and the transpersonal planet of dreams and illusion, activating the 3rd and 11th house axes and suggesting an enhanced capacity for ideation across social networks.
After (storytelling):
The way your mind works is genuinely unusual, and in a good way. When you get an idea, it does not just sit there. It travels. It finds connections across conversations, across people, across rooms. You think best out loud, in dialogue, in the middle of a group where ideas can bounce. And because your thinking has this dreamy, intuitive quality to it, you often land on something that feels ahead of its time before you can fully explain why. That is not a flaw. That is the gift.
Installation
Drop the storytelling/ folder into your Claude Code skills directory, typically:
~/.claude/skills/storytelling/
The skill is defined entirely in SKILL.md. The evals/ folder contains test cases used to measure trigger accuracy and output quality.
Repo contents
SKILL.md— the skill definition (frontmatter + instructions Claude reads at trigger time)evals/evals.json— evaluation cases for trigger accuracy and output quality