Honestly, I Never Thought "Pottery" Could Become an AI Skill
To be perfectly honest, I never imagined that a craft like "pottery making" could be turned into an AI skill. But the other day, when I stumbled upon wonsukchoi's domain-experts repository, I suddenly felt this project had real potential. This project can be summed up in a single sentence—it turns all human experts into AI agents. Just think about it: if a master potter who has spent thirty years working with ceramics could have his experience distilled into an AI character, ready to answer questions anytime, anywhere, that would be an absolute game-changer for apprentices just entering the field.
What Exactly Can a Manufacturing Potter Role Do?
Under the roles directory in domain-experts, there is a folder called manufacturing-potter, and inside it sits a SKILL.md file—that's the core of the entire skill. It packages the knowledge of a manufacturing potter into an AI expert role, and this isn't just about casually chatting over basic pottery trivia. It covers the actual, full production workflow in a very practical sense.
- Clay materials: It can provide recommendations on the characteristics of different clays, mixing ratios, and how to control moisture content.
- Forming stage: The techniques of wheel throwing, the pressure for hand-building, the timing for slip casting—these experience-based know-hows have been broken down into conversational knowledge points.
- Decoration and glazing: Glaze formulas, glaze application thickness, drying times—just ask it.
- Firing control: Kiln temperature curves, heating and cooling rates, atmosphere control, plus troubleshooting for defects like cracks and blistering after firing.
Simply put, this skill turns the "feel" in a master craftsman's hands into a knowledge base that can be queried over and over again. For a ceramics factory, it can serve as a quick reference for quality issues; for a craft course at a school, it's an on-call teaching assistant.
How to Install It in Your Own Environment
Installation is actually quite simple—it's just one command line away, provided you have Node.js set up on your computer.
npx skills add https://github.com/wonsukchoi/domain-experts --skill manufacturing-potterOnce installed, you can load this role in any AI runtime environment that supports the skills protocol. The entire role definition lives in the SKILL.md file. If you want to dig into the details, just head over to the roles/manufacturing-potter directory in the repository and take a look.
I Gave It a Try Myself
To be honest, I've only ever thrown a few crooked cups in my life, so when testing it out, I deliberately picked a few more technically challenging questions. For instance, I asked, "How do you generally adjust the ratio of ball clay to petuntse in a clay body?" The answer it gave actually made a lot of sense—it first asked about preconditions like firing temperature and forming method, then offered recommendations, rather than just reciting a manual upfront.
Later, I asked, "How do you handle pinholes on a glaze surface?" It actually broke the answer down into different scenarios: was the glaze ground too coarse, was the bisque surface not cleaned properly, or was the firing temperature increased too quickly? It walks you through the troubleshooting step by step. That kind of logical reasoning genuinely resembles a master craftsman doing on-site diagnosis—observe, question, inspect, and conclude.
A Few Tips and Things to Keep in Mind
After using it for a few days, I've summed up some lessons learned to share with you:
- Be as specific as possible about your scenario and conditions when asking. For example, "I'm using an electric kiln with a firing temperature capped at 1200 degrees Celsius"—the answers will be much more precise.
- Don't directly copy an uncertain formula. It's best to run small sample tests with your own materials. The AI provides empirical reference, but it can't fully replace experimental data.
- This skill's essence is a "role-playing" style expert, so the more you follow up and describe details, the better the results will be compared to a single isolated question.
- If you come across terminology you don't understand in the answers, just ask it to explain—it will switch into teaching mode.
How It Can Be Used in Real-World Scenarios
After thinking it over, I believe this thing has at least three practical directions worth pursuing. The first is teaching—students at arts and crafts colleges can use it as a portable craft Q&A device; flipping through reference materials during evening self-study is nowhere near as fast as just asking it directly. The second is production—when a small ceramics workshop encounters quality fluctuations, it can be used for an initial assessment, narrowing down the scope of the problem before applying targeted fixes. The third is creative assistance—when a potter is conceptualizing a new piece, they can use it to validate the feasibility of a technical route, like asking, "For this kind of irregular-shaped vessel, is slip casting or hand-building more stable?" That saves you from going down the wrong path.
Of course, at the end of the day, it's an AI role—it can't replace a real master craftsman. But as a vehicle for knowledge preservation and transfer, I genuinely think this concept is impressive. After all, if a master's experience isn't organized in a timely manner, it gradually fades away. Turning artisan wisdom into reusable skills like this means that anyone who wants to learn ceramic craft in the future suddenly has a mentor who's online around the clock.
If you're interested in ceramic craftsmanship or AI agents, I'd encourage you to browse through that repository. It doesn't just have the potter role—you might discover other fascinating expert personas in there as well. As for me, I've already bookmarked the installation command and plan to dig deeper into the SKILL.md format when I have free time, with the goal of building a role of my own.
