/make-it-repeatable

Garden · not yet reviewed Beta, from the WDAI Build Cycle pilot Version 0.1.0

Turns something that worked once into something reusable (a saved prompt, template, checklist or personal skill) plus a short share-out of what was built and learned, and can draft it as a contribution to AI Practice Commons.

How to call it

Once it's installed, ask for it by name, for example "use make-it-repeatable on this" or type /make-it-repeatable . Your AI tool also starts it without being asked when your request matches its description.

When it starts: someone says "that worked", "I'll need this again", "save this", "make this a template", "how do I do this every week", "share what I learned", or has done the same kind of task with AI more than twice.

Connections

Practises

• Make it repeatable

Concepts it uses

• Automation (garden)
• Skills (garden)
• Scheduled tasks (garden)

Comes after

• /prove-it (garden)
• practice-coach (garden)
• /build-brief (garden)
• /stuck-or-pivot (garden)

Try next

• skill-creator (garden)
• no-ai-slop (garden)

Used in blueprints

• Build a style guide so AI drafts sound like you (garden)
• Teach your AI assistant your team's jargon, one confirmed term at a time (garden)

From the field, after this

• skill-creator (garden)
• evals-skills (garden)
• compound-engineering (plugin) (garden)
• ec-public (consulting workflows) (garden)
• no-ai-slop (garden)
• Wharton Generative AI Labs prompt library (garden)

The skill itself

This is the instruction text your AI tool reads, shown as written.

Make it repeatable

Turn one good result into a routine the person can rely on.

How this skill behaves

They say what worked first; you shape it into something reusable. “Just do it” drafts the template from the conversation.

Steps

1. What worked? Ask what they did this time that they’d want to do the same way next time, and what they’d hand over to AI next time that they did by hand.
2. Capture it. Write down:
   • the inputs;
   • the request that worked (the actual prompt);
   • the steps;
   • the checks they ran, from prove-it;
   • what still needs a person. Save it as a template or checklist in a form they can paste next time.
3. Is it ready to run without them? Only suggest automating it when all of these are clear: where the inputs come from, what permissions it needs, what happens when it fails, a manual fallback, and who owns it. Otherwise keep it as a template they run by hand. Point to Anthropic’s skill-creator if they want to turn it into their own skill.
4. Share what you learned. Draft a short share-out:
   • the problem;
   • what was built;
   • what AI did, and what they checked and decided;
   • something they stopped or changed, and why;
   • the next step. They can use it in a team update, a performance review or a demo.

Give it back to the commons (optional)

If they’re willing, offer to turn what worked into a garden contribution to AI Practice Commons:

• a project card, using the commons’ project card headings;
• a new parallel or product name on a concept card;
• a resource that helped.

Strip names, workplace details and anything private first. Draft the file in the commons format. Then offer git-coach to walk them through opening the pull request themselves. Sharing is always their choice.

Output

A reusable template or checklist, a readiness note on automation, a five-line share-out, and, if they want, a draft commons contribution.

---

Credit: Demo Day share-out from WDAI's Build Cycle (programs team); automation readiness checks from the AI Practice Commons PRD.

Licence: Apache-2.0 (instructions) and CC BY 4.0 (teaching text). From AI Practice Commons by Women Defining AI; method from the WDAI Build Cycle (beta).. Plain Markdown for agents: /skills/make-it-repeatable.md