3.3 Ready-made skills | The World After Claude

💡 New here? Three words before you start

Skill — a short recipe that teaches the AI to do ONE task your way (e.g., “summarize a meeting”). It’s text, not programming.
Library — a collection of different skills saved and ready for you to pick up and use when you need them.
Adapt — take an existing skill and swap out the pieces (name, rules, tone) so it works for your case.

💎 The value of a skills library

Imagine a cook who has to come up with a new recipe every time they bake a cake. Exhausting, right? A skill it’s that written recipe: you get it right once, and your Jarvis repeats it the same way every time, without you having to explain everything again.

A skills library it’s the whole cookbook. Instead of starting from scratch, you open the shelf, pick the skill closest to what you need, and only adjusts. It’s the difference between building a piece of furniture from scratch and assembling a kit that’s already cut.

🔑 The central idea

Reusing beats reinventing — almost always:

•Without a library: every new task = explain everything to the AI all over again.
•With a library: open the shelf, grab the ready-made recipe, adjust one detail, and use it.

Each card is a ready-made recipe. Blue is where you save your next adapted skill.

A futuristic shelf lit in purple, filled with glowing skill cards ready to use — your Jarvis skill library. — Your skills shelf: the fuller it is, the less you have to reinvent.

Key concepts

📌 Skill = recipe writing

📌 Library = recipe book

📌 Reuse > reinvent

📌 Grab and adjust

⚡ Productivity Skills

Productivity skills are the the basics — the ones that save the most time on day one. Summarize a meeting, organize your inbox, turn a mess of notes into a to-do list. Nothing glamorous, all useful.

They’re perfect to start with because the result is easy to measure: either the meeting became a clear summary, or it didn’t. See what belongs in this group and what doesn’t:

✓ It’s a productivity skill

✓Summarize a meeting into bullet points and tasks
✓Inbox triage (what to respond to first)
✓Making a to-do list from scattered notes
✓Prepare the agenda for an upcoming meeting

✗ It isn’t (it’s another group)

✗Writing a social media post → that's content
✗Research a new topic in depth → that's research
✗Compare 3 vendors → that’s analysis

💡 Practical tip

Start with the annoying task you repeat every week. That’s where a ready-made skill pays off fastest: you feel the relief right away and gain confidence to adapt the next one.

Key concepts

📌 Summarize · organize · list

📌 Save time already

📌 Easy-to-measure result

📌 Start with the repeated task

✍️ Content Skills

Content skills help AI create things in your voice: posts, emails, scripts, captions, descriptions. The secret here isn’t getting AI to write—it already knows how. It’s getting it to write your way, and not in a generic robot voice.

That’s why a good content skill always keeps who you are, who you’re talking to e how you speak. With this in place, you can produce at scale without sounding like someone else.

🎙️ What makes the skill sound like you

•Tone: direct? good-humored? formal? — write that in the skill.
•Audience: who’s reading (nontechnical person, technical person, client) changes everything.
•Examples: paste 1 or 2 examples of your writing for the AI to copy your style.

One idea goes in; several formats come out — all in the same voice, because the skill carries your tone.

💡 Practical tip

Save a “voice parent skill” with 2 examples of your writing and your writing quirks. Then your post, email, and script skills just point to it. Change the voice in one place—and everything follows.

Key concepts

📌 Voice and tone set

📌 Who speaks · to whom

📌 Examples > instructions

📌 Scale without losing your voice

🔬 Research and Analysis Skills

These skills transform hours of reading in minutes: research a new topic, compare options side by side, summarize a long document, and pull out only what matters. It’s the category that impresses beginners the most.

But there’s one concern: AI can seem confident and be wrong. That’s why a good research skill always asks it to show where each thing came from. Trust without a source doesn’t count.

✓ Good research skill

✓Asks for the sources behind each claim
✓Separate “fact” from “opinion/estimate”
✓Delivers a short summary + the details
✓Alerts you when it’s unsure

✗ Dangerous research skill

✗Throws out numbers without saying where they came from
✗Mixes guesswork with facts without saying so
✗Answers confidently even in the dark

💡 Practical tip

End every research skill with a magic line: "if there isn't a reliable source, say 'I don't know' instead of making something up". That one sentence prevents most wrong answers.

Key concepts

📌 Research · compare · summarize

📌 Always ask for a source

📌 Fact ≠ guesswork

📌 "I don't know" > make it up

🪡 Adapt a ready-made skill

Here’s the trick: you almost never create a skill from scratch. You take a similar one and swap out the pieces. A well-made skill already comes with marked blanks—the <snippets between arrows> — for you to fill in with your reality.

Adapting means replacing the name, the rules e o tone. Below is a real skill — “Summarize meeting” — ready to copy, paste into Claude or ChatGPT, and use today. Change only the parts between < >.

📥 Grab a Similar Skill

Choose the one that comes closest to what you want — not the perfect one.

🪡 Replace the <sections>

Fill in the blanks with your name, your audience, and your rules.

✅ Test and adjust

Run it once, see what went wrong, and fix the skill text.

🧪 Ready-made skill: "Summarize a meeting" (copy and use)

Goal: turn messy meeting notes into a clear summary with decisions and tasks. Paste it into Claude or ChatGPT and replace the parts between < >.

Você é meu assistente de reuniões.

Vou colar minhas anotações da reunião sobre <assunto da reunião>.
Os participantes foram <nomes, ex: eu, João e a Ana>.

Me devolva, nesta ordem:
1. RESUMO em até 3 frases.
2. DECISÕES tomadas (lista com bullets).
3. TAREFAS: o que ficou pra cada pessoa, com prazo se houver.
4. PENDÊNCIAS: o que ficou em aberto.

Regras:
- Use linguagem simples, no tom <formal ou informal>.
- Se algo não ficou claro nas notas, marque com "⚠️ confirmar".
- Não invente decisão que não está nas notas.

Minhas notas: <cole aqui suas anotações>

How to verify it worked: the answer must have the 4 sections in order, each task with an owner, and a "⚠️ confirm" whenever the note was vague—without inventing decisions.

Key concepts

📌 Adapt > build from scratch

📌 <snippets> = your gaps

📌 Change the name · rule · tone

📌 Test, review, adjust

🧭 Where to find more skills

The good news to wrap up: you never runs out of raw material. Ready-made skills are everywhere—you just need to know where to look. And the more you collect, the richer your shelf becomes.

👥

Communities: AI groups share skills all the time—copy the ones that work for you.

📦

Repositories: open collections on the internet with hundreds of skills organized by topic.

🎓

INEMA courses: each course already provides tested, ready-to-use skills — a gold mine for your Jarvis.

🧰

You do it yourself: every time you adapt a skill, save it—your shelf grows on its own.

🎯 The habit that makes a difference

Whenever you find a good skill, save it before you need it. A full shelf is what separates people who struggle with AI every time from those who just grab a recipe and get on with life.

✋ Before you continue — what's the best way to use the skills library?

🎓 Module summary

✓

A skill is a written recipe — and the library is the cookbook you can reuse.

✓

Three useful families — productivity, content, and research/analysis.

✓

Adapt > create from scratch — replace the <snippets>, test, and adjust.

✓

Skills are everywhere — INEMA communities, repositories, and courses.

Next module:

3.4 — 🧱 Ready-made agents and workflows: templates for agents and multi-step flows for you to adapt.

← Back to the Track Next: 3.4 Ready-made agents and workflows →