simota-agent-skills — Claude Code skill: design suite and setup
simota-agent-skills — a collection of design skills for interface direction, Figma context, mockup-based HTML/CSS, and UX copy. Call a focused skill or let Vision coordinate design work.
MIT 199 installs ★ 35 updated 2026-10-21
npx skills add podo/design-agent-skills --skill simota-agent-skills -a claude-codeWhat simota-agent-skills does
The suite includes vision for UI/UX direction and design systems, muse for tokens and dark mode, frame for Figma context, and pixel for producing HTML/CSS from mockups and visually checking it. prose handles user-facing text; ink makes SVG icons and illustrations. echo simulates user perspectives to inspect flows, but it is not research with real users.
Each skill can be invoked on its own, while vision can coordinate the other design skills. frame uses the Figma MCP server to extract design context. With the installation route described for this collection, shared files from _common are not copied, though each skill's own reference/ folder is installed. Some shared instructions are therefore unavailable.
There is a name clash between ink and ink-google: installing both overwrites the one installed earlier. The collection also lists Codex, Cursor, GitHub Copilot, Gemini CLI, and OpenCode as platforms.
What it helps with
- build a project design system
- prepare Figma context for interface implementation
- create HTML/CSS from a mockup image
- inspect a user flow with simulated personas
- write onboarding and error messages
Example prompts for the agent
- Use
visionto shape the design system for this project and coordinate the design work. - Use
frameto extract the Figma context needed to implement this interface. - Use
echoto review this flow from simulated user perspectives and identify unmet needs.
How to install simota-agent-skills
1Skills CLI (Vercel)
npx skills add podo/design-agent-skills --skill simota-agent-skills -a claude-code2Manually: put the files in place
~/.claude/skills/simota-agent-skills/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/simota-agent-skills/
1Skills CLI (Vercel)
npx skills add podo/design-agent-skills --skill simota-agent-skills -a codex2Manually: put the files in place
.agents/skills/simota-agent-skills/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add podo/design-agent-skills --skill simota-agent-skills -a cursor2Manually: put the files in place
.cursor/skills/simota-agent-skills/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install podo/design-agent-skills skills/simota-agent-skills2Manually: put the files in place
.github/skills/simota-agent-skills/SKILL.md1Skills CLI (Vercel)
npx skills add podo/design-agent-skills --skill simota-agent-skills -a gemini-cli2Platform CLI
gemini skills install https://github.com/podo/design-agent-skills --path skills/simota-agent-skillsManually: put the files in place
.devin/skills/simota-agent-skills/SKILL.mdThe legacy .windsurf/skills/ path still works
Skills CLI (Vercel)
npx skills add podo/design-agent-skills --skill simota-agent-skills -a opencodeClaude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
Step by step, with commands for each platform: How to install skills in Claude Code: five ways
Questions about simota-agent-skills
How do I install simota-agent-skills in Claude Code?
Run npx skills add podo/design-agent-skills --skill simota-agent-skills -a claude-code. The skill file is placed at ~/.claude/skills/simota-agent-skills/SKILL.md.
What does simota-agent-skills do?
It provides separate skills for interface design, design tokens, Figma context, mockup-based markup, UX copy, and related work. vision can coordinate the design-focused skills.
Does simota-agent-skills need a Figma MCP server?
The frame skill uses a Figma MCP server to extract design context. The other skills cover different design tasks.
Source text
SKILL.md · MIT
--- name: simota-agent-skills description: | Large collection of specialist agent skills coordinated by an orchestrator, with a design suite: Vision (creative direction, design systems), Muse (design tokens, theming, dark mode), Frame (Figma-to-code context via MCP), Pixel (pixel-accurate HTML/CSS from mockups), Echo (simulated persona UX validation), Prose (microcopy and UX writing), and Ink (SVG icons and illustrations). The catalogue installs only these design skills. triggers: - "simota" - "Vision skill" - "Muse skill" - "Frame skill" - "Pixel skill" - "design suite" das: type: package category: interaction-polish upstream: "https://github.com/simota/agent-skills" version: latest install: false --- # simota-agent-skills > Catalogue stub — full package: [simota/agent-skills](https://github.com/simota/agent-skills) <!-- das:install:start — generated from stub.yaml by `npm run generate`; edit stub.yaml, not this block --> ## Decision tree 1. **Is the full skill already installed?** - Global: `f=~/.agents/skills/vision/SKILL.md; [ -e "$f" ] && ! grep -qs "^das:" "$f" && echo "installed" || echo "pointer"` - Project: `f=.agents/skills/vision/SKILL.md; [ -e "$f" ] && ! grep -qs "^das:" "$f" && echo "installed" || echo "pointer"` - `installed` → invoke it and proceed - `pointer` → go to step 2 2. **Install** (the skills CLI cannot install this one): The skills CLI cannot copy this repo as-is, because every skill folder links to a shared _common folder that links back (cyclic symlinks, ELOOP), and it reports success while installing nothing usable for Claude Code. A sparse clone of only the design-related skill folders leaves those links dangling, so the CLI installs each skill with its reference folder but without the shared _common docs. Upstream's own route is `make link` from a full clone, which symlinks every skill in the repo into ~/.claude/skills, ~/.agents/skills and ~/.gemini/antigravity-cli/skills. If an earlier attempt left `/tmp/simota-agent-skills` behind, delete it first (`rm -rf /tmp/simota-agent-skills`): `git clone` refuses a non-empty directory. ```bash [ -e ~/.agents/skills/simota-agent-skills ] && echo "global" || echo "project" ``` **Global:** ```bash git clone --depth=1 --filter=blob:none --sparse https://github.com/simota/agent-skills.git /tmp/simota-agent-skills git -C /tmp/simota-agent-skills sparse-checkout set vision muse frame pixel echo prose ink npx skills add /tmp/simota-agent-skills --skill vision muse frame pixel echo prose ink -g -y ``` **Project:** ```bash git clone --depth=1 --filter=blob:none --sparse https://github.com/simota/agent-skills.git /tmp/simota-agent-skills git -C /tmp/simota-agent-skills sparse-checkout set vision muse frame pixel echo prose ink npx skills add /tmp/simota-agent-skills --skill vision muse frame pixel echo prose ink -y ``` **Heads-up** - **License:** MIT - **Name collision:** `ink` is also written by `ink-google` (its install or its catalogue pointer); installing both overwrites the earlier one. <!-- das:install:end --> ## Invoke after install - Skill names: `vision`, `muse`, `frame`, `pixel`, `echo`, `prose`, `ink` - Trigger phrases: "simota", "Vision skill", "Muse skill", "Frame skill", "Pixel skill", "design suite" ## What it does The design-related members of a large agent roster, each independently invocable: - `vision` — directing UI/UX creative work: redesigns, new designs, trend application, design-system construction; orchestrates the other design agents and has a co-design pair mode - `muse` — defining and managing design tokens, applying token systems to existing codebases, dark mode, cross-platform output - `frame` — extracting and structuring design context from Figma through the Figma MCP server for downstream implementation, or Code Connect management - `pixel` — generating pixel-accurate HTML/CSS from image mockups and checking the result visually - `echo` — simulating users to evaluate existing flows: cognitive walkthroughs, unmet needs, JTBD, opportunity trees (not real-user research) - `prose` — user-facing UX text: microcopy, error messages, voice and tone, onboarding copy, accessibility text - `ink` — SVG icons, icon systems, sprite symbols, and illustrations The skills refer to each other and to shared protocol docs in the repo's `_common` folder; the install route used here cannot copy `_common` (see the install section), so those shared docs are absent, while each skill's own `reference/` folder is installed. The repo has many more agents, including design-adjacent ones (`palette` for usability and accessibility, `flow` for CSS/JS animation, `artisan` for production frontend code, `vitrine` for Storybook, `atelier` for design-to-implementation pipelines, `forge` for prototypes, `canvas` for diagrams). To add one, put its folder name in both the `sparse-checkout set` command and the `--skill` list. ## When NOT to use - When you need a single focused skill — individual skills from this suite can be preferred over installing all seven - If you only need design review without the full toolkit — use `plan-design-review` or `creative-director` instead - The generic folder names of the wider roster (`launch`, `schema`, `matrix`, `pulse`, ...) collide with unrelated skills, so do not install the whole repo
© podo · MIT · podo/design-agent-skills