design-review-garrytan
Skill32 starsChanged 6 months agoMITView on GitHub
Designer Who Codes: visual audit followed by atomic commits fixing each issue. Garry Tan's design review skill — treats UI like code, identifies and fixes design debt systematically.
In this repository170 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- accessibility-agentsSkill
- accessibility-catalogueSkill
- addyosmani-qualitySkill
- ai-graphic-design-skillSkill
- algorithmic-artSkill
- anatomySkill
- animate-css-skillSkill
- animate-skillSkill
- anthropics-skillsSkill
- antvis-chart-skillsSkill
- apple-hig-skillsSkill
- assimovt-productskillsSkill
- awesome-design-skillsSkill
- baseline-uiSkill
- bencium-ux-designerSkill
- brand-design-mdSkill
- building-componentsSkill
- callstack-agent-skillsSkill
- cc-slidevSkill
- chadboyda-gtmSkill
- claud3Skill
- claude2figmaSkill
- claudedesignskillsSkill
- claude-pm-skillsSkill
- claude-wireframe-skillSkill
- cloudai-threejsSkill
- cloudflare-web-perfSkill
- color-expertSkill
- composio-artifactsSkill
- content-catalogueSkill
- content-strategySkill
- copywriting-skillSkill
- coreyhaines-marketingSkill
- creative-directorSkill
- css-animation-skillSkill
- d3js-skillSkill
- dark-pattern-auditSkill
- data-analysis-skillSkill
- data-viz-agentSkill
- deanpeters-pm-skillsSkill
- design-auditorSkill
- design-briefSkill
- design-catalogueSkill
- design-consultationSkill
- design-debt-auditSkill
- design-engineering-catalogueSkill
- designer-skillsSkill
- design-for-aiSkill
- design-htmlSkill
- design-impact-reportingSkill
- design-labSkill
- design-motion-principlesSkill
- design-negotiationSkill
- design-sprintSkill
- design-system-governanceSkill
- design-tokens-skillSkill
- design-with-claudeSkill
- digidai-pmSkill
- dimensional-drawingSkill
- The first 60 shown.
The file
Files
design-review-garrytan
Catalogue stub — part of garrytan/gstack (
design-review)
<!-- das:install:start — generated from stub.yaml by `npm run generate`; edit stub.yaml, not this block -->Decision tree
- Is the full skill already installed?
f=~/.claude/skills/gstack/design-review/SKILL.md; [ -e "$f" ] && ! grep -qs "^das:" "$f" && echo "installed" || echo "pointer"installed→ invoke it and proceedpointer→ go to step 2
Install (the skills CLI cannot install this one):
This skill is part of gstack and calls gstack's shared tooling (
~/.claude/skills/gstack/bin/…), so install gstack itself rather than the single skill (--skill design-reviewfails because the CLI only sees gstack's root router, and even with--full-depththe copied instructions lack the tooling)../setupneeds Bun, builds binaries, downloads Chromium and registers a Stop hook in your Claude settings (opt out with./setup --no-timeline-stop-hook); ask the user before running it. Upstream's README also tells agents to add a gstack section to CLAUDE.md after setup.git clone --single-branch --depth 1 https://github.com/garrytan/gstack.git ~/.claude/skills/gstack cd ~/.claude/skills/gstack && ./setup
Heads-up
- Requires: Bun (setup refuses versions older than 1.3.3; upstream tests 1.4.2+), Git and Claude Code
- License: MIT
- Download size: ~98 MB.
<!-- das:install:end -->Invoke after install
- Skill name:
design-review - Trigger phrases: "design review", "visual audit", "UI audit", "fix design issues"
What it does
"Designer Who Codes" approach: performs a structured visual audit of the UI, then fixes each issue with an atomic commit. Treats design debt like technical debt — identifies it, prioritizes it, and ships fixes. From Garry Tan's gstack, his toolkit for building well-designed products.
It expects a clean working tree (it offers to commit or stash your changes first) because every fix lands as its own commit.
When NOT to use
- Planning a design before anything is built — use
plan-design-reviewordesign-consultation - A review that must not touch code or git history — this skill edits files and commits
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

