Where to Go Next

You built a skill. Here is the path most teams follow from there.

Read the article series

The Build Your Software Factory articles go deeper into the practices you just touched on. They are short, hands-on, and build on each other.

Article 1 — Do the Work, Then Capture It — The full version of the loop you just ran, with worked corrections.
Article 2 — Test Your Skill in a Clean Room — How to verify a skill works without leaning on context the assistant should not have.
Article 3 — The Retrospective: “Did You Use a Skill?” — The single team habit that turns one-off prompts into durable factory components.
Article 4 — Sharpen Your Skill with /skill-optimizer — Apply authoring best practices to a skill you already have.
Article 5 — A Skill That Holds Up Under Pressure — What it looks like when a skill graduates from “works on my machine” to a team standard.

Browse the full list on the Articles page.

Set up your IDE

Each AI IDE has its own folder conventions and sharing requirements. The IDE guides cover everything you need — skills, commands, agents, workflows — for the tool you actually use.

If your team uses more than one IDE, see the sharing strategies on the Skills page.

Explore example factories

Working factories are the fastest way to develop intuition for what good composition looks like. Each example pairs a real application style with the skills, commands, agents, and workflows that support it.

Browse example factories →

Pick one close to what your team builds, read its components, and steal what fits.

Add the next skill

The first skill you built is the proof of concept. The second, third, and fourth are where the factory starts paying for itself. Same loop every time:

Do a task you do often.
Correct the output until it meets your standards.
Ask the assistant to capture the standards as a skill.
Test it in a fresh session.
Commit and share.

Repeat until the assistant produces team-quality output on the tasks that matter to you.