Open dotted notebook with blue-ink dashboard wireframe sketches and handwritten design notes, a wooden pen resting on the page
Open dotted notebook with blue-ink dashboard wireframe sketches and handwritten design notes, a wooden pen resting on the page

•

9 mins read time

Gemini Skills for Designers: Turn Your Design Process Into Reusable AI Tools

Google released skills in Gemini on 30 September 2026, replacing Gems with reusable, stackable instructions that can carry reference files. Here is what Gemini skills mean for designers, five worth building, where the hype outruns reality, and whether to worry.

Ayoub Kada

•

9 mins read time

Gemini Skills for Designers: Turn Your Design Process Into Reusable AI Tools

Google released skills in Gemini on 30 September 2026, replacing Gems with reusable, stackable instructions that can carry reference files. Here is what Gemini skills mean for designers, five worth building, where the hype outruns reality, and whether to worry.

Ayoub Kada

A skill makes Gemini consistent. It does not make Gemini right. The value is in how well you write down how you work.

Gemini skills for designers: the short answer

Google released skills in Gemini on 30 September 2026. According to Google's announcement of skills in Gemini, a skill is a saved set of instructions you create once, call by typing a forward slash and its name, and reuse in any chat. Skills can include reference files such as text documents, PDFs and images, several can be stacked in one conversation, and Gemini can run one automatically when your prompt matches it. Skills are set to replace Gems.

What Gemini skills for designers really means is this: the instructions you keep retyping (your critique checklist, your UX writing voice, your accessibility review, your design system rules) can become small, named tools that travel with you. That is a real improvement in how designers work with AI. It is not a new design capability. A skill makes Gemini consistent. It does not make Gemini right.

What Google actually shipped

Kept to what the announcement states:

Reusable instructions. You can turn a chat into a skill and run it again with a slash command, instead of pasting the same long prompt every time.

Reference files inside the skill. A skill can carry plain text documents, PDFs or images. For designers, that is the important part: a skill can hold your design principles, a brand voice PDF or a screenshot of a reference pattern.

Stacking and automatic use. More than one skill can apply to the same chat, and Gemini can trigger a skill on its own when a prompt matches.

Availability. Google says skills are rolling out globally in Gemini chat for users aged 18 and over, with Google Workspace business, enterprise, nonprofit and education customers following in the coming weeks.

Gems are being retired. Existing Gems will be migrated to skills automatically. Google lists support ending for Gems in November 2026 for personal accounts, March 2027 for Workspace business, enterprise and nonprofit, and June 2027 for Workspace education.

That is the release. No design canvas, no Figma integration announced with it. The change is in how you package your own expertise for the model.

Hype vs reality

The hype version is that skills turn Gemini into "your own design assistant trained on your process". Partly true, mostly framing.

What holds up: repetition was the quiet tax on using AI for design work. Every review started with the same paragraph of context, and every result drifted because the context was slightly different each time. A saved skill with attached reference files removes that drift. Two designers on the same team calling the same critique skill get feedback against the same criteria. That is worth having.

What does not hold up: a skill is a prompt with a name and some files. It is not training, it does not learn your taste over time, and it does not know anything you did not write down. If your design principles are vague, the skill repeats the vagueness on demand. Automatic triggering also cuts both ways: a skill that fires when you did not expect it can quietly shape an answer you thought was neutral.

The verdict: Gemini skills are a genuine workflow upgrade and slightly overhyped as "personal AI". Their value depends entirely on how well you can write down how you work.

Gems vs skills vs pasted prompts

Approach

How you use it

Reference material

Best for

Main weakness

Pasted prompt

Copy and paste each time

Whatever you remember to attach

One-off questions

Drifts every time it is retyped

Gem (being retired)

Pick a Gem before the chat starts

Knowledge files

A dedicated assistant for one job

One configuration per session

Gemini skill

Type / and the name, or let Gemini trigger it

Text, PDFs and images

Repeatable design tasks inside any chat

Only as good as the instructions written

Tool-specific skill (for example in a design tool)

Lives inside the product

Project pages, components, styles

Work that must follow one project's system

Locked to that tool

The last row is worth noting. Framer added Skills for the Framer agent on 22 September 2026, letting teams save design system and CMS instructions. The pattern is spreading: across tools, the unit of AI customization is becoming the named, reusable skill. Designers who learn to write good ones will use that skill everywhere.

Five Gemini skills worth building as a designer

These are suggestions for what to build, not results from testing. Each one replaces a prompt designers commonly retype.

/critique. Your heuristics for reviewing a screen: hierarchy, clarity of the primary action, empty and error states, consistency with the system. Attach your design principles document. Ask it to return issues ranked by impact, not a list of compliments.

/ux-copy. Your product's voice and terminology. Attach a short PDF of voice rules and a glossary of words you use and avoid. Use it for button labels, error messages and onboarding text.

/a11y-check. An accessibility review pass on a screenshot: contrast concerns, focus order questions, touch target size, text alternatives. It flags what to verify. It does not replace testing with real tools and real users.

/research-synth. How you want interview notes summarized: separate observation from interpretation, quote participants accurately, list open questions. This is the one where a stable format matters most.

/handoff-notes. The structure your engineers want for a feature spec: states, edge cases, motion notes, analytics events. Attach an example of a good past spec.

Stacking is where this gets useful. Running /critique and /ux-copy together on the same screenshot gives a structural review and a copy review in one pass, each against its own written rules.

The Trigger, Rules, References, Output template

A simple structure for writing a skill that holds up. Call it the Trigger, Rules, References, Output template.

Trigger. One sentence on when this skill should run. Because Gemini can invoke skills automatically, a precise trigger ("when I share a mobile screen and ask for a review") prevents it firing on unrelated chats.

Rules. The criteria, written as checkable statements. "Primary action is visually dominant on the screen" is checkable. "Make it feel premium" is not.

References. Attach the files the rules depend on: principles, voice guide, a token sheet, one good example. Keep them short. A long PDF the model skims is weaker than a one-page summary it reads fully.

Output. The exact format you want back: a ranked list, a table, a rewritten version beside the original. A fixed format makes results comparable across people and over time.

Write the skill, run it on three real screens you already know well, and fix the rules where the output disappoints you. That loop is where the actual value is created.

Where design work is going

Skills are part of a bigger direction: design knowledge is moving out of people's heads and slide decks into structured instructions that AI tools read. The same thing is happening in code, where the Claude Sonnet 5.5 changes for designers depend on how clearly you write design system rules, and in interfaces, covered in the piece on generative UI and interfaces that assemble themselves.

Over the next one to three years, expect three shifts. Teams will maintain libraries of shared skills the way they maintain component libraries today, with owners and versions. Writing precise, testable design criteria becomes a core design skill, not an afterthought. And reviews get faster but not shorter, because the time saved on routine checks moves to the decisions no skill can make.

The fear: will Gemini skills replace UX designers?

No. But the fear behind the question deserves a straight answer.

Skills make it easy to package a checklist. If a large part of your job is applying a checklist (routine reviews, standard copy, formatted specs), that part gets faster for everyone, including people who are not designers. Pretending otherwise does not help.

What a skill cannot package is the judgment that wrote the checklist: knowing which rule to break for this user, noticing the problem nobody asked about, deciding what the product should be. The longer argument is in will AI replace UX designers. The short version: designers who can encode their thinking into tools gain leverage. Designers who only follow encoded rules lose it.

Concrete moves for this month:

  1. Write down your three most repeated prompts and turn them into skills using the template above.

  2. Rewrite one vague principle in your design system as a checkable rule.

  3. Share one skill with a teammate and compare results on the same screen. Disagreements show where your rules are unclear.

  4. If you rely on Gems, note the migration dates and check your Gems after they convert.

When skills are the wrong tool

Use Gemini skills for repeatable, text and image based tasks with stable criteria. Do not use them for decisions that need context the skill cannot hold: product strategy, research with real users, or anything regulated where an automatic, unseen instruction is a risk. And if you are designing an AI product of your own, the hard part is the UX of the AI itself, which no prompt template solves.

For teams building AI features into their products, DIGCY offers AI product development and UI/UX design services. DIGCY is a Dribbble Selected Agency and a UI/UX and product design studio in Morocco working with clients globally.

FAQ

What are skills in Gemini?

Skills are reusable custom instructions in Gemini, released by Google on 30 September 2026. You save instructions once, call them by typing a forward slash and the skill name, attach reference files such as PDFs or images, and stack several in one chat. Gemini can also run a skill automatically when a prompt matches.

How can designers use Gemini skills?

Designers can turn repeated prompts into skills: a design critique checklist, a UX writing voice guide, an accessibility review pass, a research synthesis format or a handoff spec structure. Attaching principles or voice documents as reference files keeps results consistent across a team.

What is the difference between Gems and skills in Gemini?

A Gem is selected before a chat starts and configures that session. A skill is called inside any chat with a slash command, can be combined with other skills and can be triggered automatically. Google is migrating Gems to skills and ending Gem support from November 2026 for personal accounts.

Will Gemini skills replace UX designers?

No. Skills automate the repeatable parts of design work, such as running a checklist or formatting a spec. They cannot frame problems, research users or decide what a product should do, and they only apply the judgment a designer writes into them.

Are Gemini skills overhyped?

Partly. The consistency gain from saved instructions and attached reference files is real. Calling them a personal AI that knows your process overstates it: a skill is a named prompt with files, and it is only as good as the rules written into it.

Let’s keep in touch.

Discover more about high-performance web design. Follow us on Twitter and Instagram.

Have a project in mind?

By submitting, you agree to our Terms and Privacy Policy.

Let's Talk

Share your project requirements here or send us an email at hello@digcy.com we will follow up in less than 24 hrs

Have a project in mind?

By submitting, you agree to our Terms and Privacy Policy.

Let's Talk

Share your project requirements here or send us an email at hello@digcy.com we will follow up in less than 24 hrs

Have a project in mind?

By submitting, you agree to our Terms and Privacy Policy.

Let's Talk

Share your project requirements here or send us an email at hello@digcy.com we will follow up in less than 24 hrs