Figma Agent Skills: My Reusable Brand Audit Command (2026)

If you’ve ever had to explain your agency’s brand rules to a new designer—or an AI—for the tenth time, this is for you. You can now write your design system conventions once, in plain English, and have anyone on your team apply them with a single slash command.

Figma just shipped Agent Skills. It lets you encode your agency’s institutional knowledge into reusable commands. This post is the how-to. I’m going to give you the exact Brand Audit skill I use, so you can stop policing your design files and let the agent do it for you.

---

What are Figma Agent Skills?

Figma Agent Skills are reusable sets of instructions, written in plain English as Markdown (.md) files, that guide Figma’s AI agent to perform specific, multi-step tasks. Instead of re-writing complex prompts, teams can publish a Skill once and any designer can trigger it with a simple slash command, ensuring consistent operations like brand audits or design reviews.

---

The Brand Audit Skill I Use — Copy It Right Now

This is the plain-text asset that powers a custom slash command in Figma. I save this as a Markdown file (e.g., `brand-audit.md`) and publish it to my team’s library. Now, any designer can open a Figma file, type `/brand-audit`, and the agent will run these checks automatically, flagging any design elements that don’t match our official tokens.

# Brand Audit Skill

> This skill audits the current Figma page against our agency's core brand guidelines. It checks for compliance with color tokens, font tokens, and spacing rules. It will flag any layers that violate these rules and generate a summary report.

## Task 1: Color Token Compliance

- Scan all visible layers on the current page.
- Compare the fill and stroke colors of every shape and text layer against the published `[Brand] Color Tokens` library.
- Flag any layer using a hex code that is not a defined token.
- List the names of non-compliant layers.

## Task 2: Font Token Compliance

- Scan all text layers on the current page.
- Compare the font family, weight, and size against the published `[Brand] Typography Tokens` library.
- Flag any text layer that uses a font style not defined as a token.
- List the names of non-compliant layers and their incorrect font properties.

## Task 3: Spacing and Grid Compliance

- Check for elements positioned off the established 8px grid.
- Measure the padding and margin between major container elements.
- Flag any spacing that is not a multiple of 8px.
- List the names of misaligned layers.

## Output

- Generate a new page in the file named "Brand Audit Report - [Date]".
- On that page, create a simple list summarizing the findings from all tasks.
- For each non-compliant layer found, add a comment directly on that layer in the original page with details of the violation.

This is the core of the talk-to-build stack applied to design operations. You turn your process into a tool that works at the speed of typing.

+----------------------+       +----------------+       +--------------------+
| Designer types       | ----> | Agent runs     | ----> | Report is generated|
| `/brand-audit`       |       | Skill (.md)    |       | inside Figma file  |
+----------------------+       +----------------+       +--------------------+

---

Here’s Exactly How I Create and Publish a Skill

My process for creating a new Figma Agent Skill involves four simple steps that take it from a plain text file on my computer to a shared command my entire agency can use. The workflow is designed for designers, not developers, and requires no code. It’s all about clearly writing down the steps you already take manually.

1. Create a Markdown File. I open any plain text editor, like VS Code or even TextEdit, and create a new file. I save it with a descriptive name and the `.md` extension, for example, `ux-writing-check.md`.
2. Write the Instructions. I structure the file like the asset above. I use a main heading (`#`) for the skill’s name, a blockquote (`>`) for a short description, and subheadings (`##`) for each major task the agent needs to perform. Under each task, I use bullet points to list the specific actions.
3. Test the Skill Locally. Before sharing, I use the Figma agent’s development mode to load the local `.md` file. I run the command in a test file to make sure the agent understands the instructions and performs the tasks correctly.
4. Publish to the Team Library. Once I’ve confirmed the skill works as expected, I publish it to our shared team library in Figma. It immediately becomes available as a new slash command to everyone in the organization, just like a shared component.

---

What This Changes for Agency Design Systems

Figma Agent Skills fundamentally change how design systems are maintained and enforced within an agency, shifting the burden from manual human review to automated AI audits. This move makes brand consistency more scalable and frees up senior designers to focus on creative problems instead of policing pixel-level compliance on every project.

Dimension	Old Way (Manual Checks)	New Way (Agent Skills)
Consistency	Senior designer manually reviews files, leaving comments. It’s slow and error-prone.	Any designer runs `/brand-audit`. The AI flags every deviation from tokens instantly.
Onboarding	New hires spend hours learning the 50-page brand guide PDF.	New hires learn a few slash commands (`/brand-audit`, `/check-accessibility`).
Design Reviews	Feedback focuses on low-level mistakes: wrong color, incorrect spacing.	The AI handles the low-level checks. Human feedback focuses on strategy and UX.
Scalability	Maintaining standards across 10+ active projects requires a dedicated design ops role.	A single set of published Skills maintains standards across hundreds of projects.

The job of a systems designer shifts from being a gatekeeper to being a tool-builder. You’re not just creating the rules; you’re creating the agent that enforces them.

---

My $0.02 — How I’d Roll This Out in an Agency

My approach to integrating Figma Agent Skills in an agency would be gradual, focusing on solving the most repetitive problems first to build trust and demonstrate immediate value. It’s a three-day sprint to go from concept to a core set of shared commands that make everyone’s life easier.

Day 1 — Build the first utility Skill. I’d identify the single most annoying, repetitive check our team does. For most, it’s the brand audit. I would write the `brand-audit.md` Skill, test it on a few recent projects, and refine the instructions until it reliably catches the top 80% of common mistakes.

Day 2 — Pilot with a friendly user. I’d grab one other designer on the team and have them use the Skill on their active project. I wouldn’t just send the file; I’d sit with them and watch. Their confusion is my homework. I’d use their feedback to clarify the instructions in the Markdown file and simplify the output.

Day 3 — Publish and announce. Once the first Skill is solid, I’d publish it to the team library and do a quick 15-minute demo in our team meeting. I’d frame it as a tool to save them time, not to police their work. I’d then immediately ask: “What’s the next boring thing we should automate?” and build the backlog for Skills #2 and #3.

---

FAQ

What are Figma Agent Skills?
Figma Agent Skills are reusable instructions written in simple Markdown files. They allow teams to encode their specific workflows, like brand audits or accessibility checks, into custom slash commands that any team member can run within Figma to ensure consistency.

Are Figma Agent Skills free to use?
The feature is free during the beta period, but it requires a paid Figma plan (Professional, Organization, or Enterprise) to create and use team-published Skills. The free Starter plan is excluded from this functionality.

Do I need to be a developer to create a Skill?
No. Skills are written in plain English using Markdown, a simple text formatting syntax. If you can write a bulleted list, you can write a Skill. It’s about clearly describing a process, not writing code.

How is a Skill different from a Figma plugin?
A plugin typically provides a user interface with buttons and panels that you operate manually. A Skill is conversational; you tell the AI what you want to achieve with a text command, and the agent uses Figma’s native tools to execute the workflow for you.

Can a Skill make changes to my design file?
Yes, depending on how you write the instructions. A skill can be configured to generate new pages, add comments to specific layers, organize your file, or even modify properties of objects, all based on the steps you define in the Markdown file.

Where do I save the .md file for a Figma Skill?
You create the .md file on your local machine. Once you’ve tested it, you publish it through Figma to your team’s shared library. This makes the associated slash command available to everyone with access to that library.

Can I use this on the Figma starter plan?
No, creating and using shared Agent Skills is a feature limited to Figma’s paid tiers: Professional, Organization, and Enterprise. The functionality is designed for team-based workflows which are the focus of these plans.

---

Want help applying this?

Four ways to go deeper:

• Build with Builders. Join the Talk-to-Build community to learn to build AI-native websites, cinematic AI video, and agent-driven workflows you can sell.
• 1-on-1 working session. Book a screen-share with me — bring a real problem, leave with a working piece of it.
• Done-for-you. MK-Way builds AEO-ready websites, apps, and AI agent workflows.
• Quick question. DM me on Instagram or LinkedIn. I read every message.

---

Part of the AI Pulse series. If you commented “FIGMA” on one of my videos — this is the breakdown.

Last updated: 2026-07-09.