There are a few moments in every designer's workflow where the work isn't design. It's wrangling. Structuring research so a PM will actually read it. Checking a Figma file for the gaps you already know an engineer will find. Turning a vague ticket into something you can actually design against.

These are Claude Code commands I built for those moments. Each one is a single markdown file. No plugins, no config, no accounts. Drop them in ~/.claude/commands/ and they're ready.

Subscribe below to get all three, plus every new skill I add.

Get the skills and every future addition.

Turn a ticket into a design brief before you open Figma

You get a PRD, a Slack thread, or a one-liner in a ticket. The expectation is that you'll "just start designing." But you know the first hour will be spent figuring out what's actually being asked. What's in scope, what states to cover, who to ask about the parts that don't make sense.

/brief reads the source material and generates a structured design brief: problem statement, user context, success criteria, scope boundaries, key flows, edge cases, and open questions routed to the right person. It doesn't design the solution. It makes sure you're solving the right problem before you start.

If the input is thin, it says so. A brief that flags gaps is more useful than one that papers over them.


Review your own design the way an engineer will

Every designer has shipped a file that started a "quick question" thread. Missing hover state. Hardcoded spacing. A layer named Group 47. You knew you should've checked. You just didn't have a checklist honest enough to catch it.

/figma-crit pulls your Figma file and reviews it through the lens of the person who has to build it. It checks spacing and layout values against tokens, flags missing interactive states, catches colors and type not mapped to design system variables, and calls out generic layer names.

It only flags blockers. If something's clear, it skips it. You get a short list of gaps. Fix them before the thread starts.


From interview notes to structured findings

You ran the sessions and took the notes. Now you're staring at a doc trying to turn 6 interviews into something a PM will act on. So you spend an afternoon formatting, re-reading, and organizing, and half the findings still end up in the wrong team's backlog.

/research takes your raw notes and synthesizes them into structured findings. Each insight gets an observation, an implication, a recommendation, a confidence level, and a team routing. It doesn't write your insights for you. It structures the mess so you can focus on the "so what."


How to install

  1. Download the .md files
  2. Drop them in ~/.claude/commands/
  3. Open Claude Code and type /brief, /figma-crit, or /research

That's it. Each file is plain markdown. Read it, fork it, customize it for your team's design system and vocabulary. They're meant to be modified.


Get the skills and every new one I ship.