Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add gabelul/stitch-kit --skill stitch-mcp-create-design-system-from-design-mdgit clone --depth 1 https://github.com/gabelul/stitch-kitWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/gabelul/stitch-kit/stitch-mcp-create-design-system-from-design-md)<a href="https://agentmods.dev/skills/gabelul/stitch-kit/stitch-mcp-create-design-system-from-design-md"><img src="https://agentmods.dev/badge/skills/gabelul/stitch-kit/stitch-mcp-create-design-system-from-design-md/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/gabelul/stitch-kit/stitch-mcp-create-design-system-from-design-md"><img src="https://agentmods.dev/badge/skills/gabelul/stitch-kit/stitch-mcp-create-design-system-from-design-md.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00046 | $0.00975 |
| Opus 5 | $0.00023 | $0.00487 |
| Sonnet 5 | $0.00009 | $0.00195 |
| Haiku 4.5 | $0.00005 | $0.00097 |
Grade A, and why
stitch-mcp-create-design-system-from-design-md scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 12d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 91 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Stitch MCP — Create Design System from DESIGN.md
Turns an uploaded DESIGN.md into a real Stitch Design System and displays it in the UI. This is step two of a pair — stitch-mcp-upload-design-md must have run first, because this tool consumes the screen instance that upload created.
Calling it without a prior upload fails: there's no screen instance to point at.
Critical prerequisite
Only use this skill when the user explicitly mentions "Stitch" in their request. Never trigger Stitch operations silently during regular conversation.
When to use
- Immediately after
stitch-mcp-upload-design-mdreturns — always, as a pair - Never on its own
Step 1: Confirm you have the screen instance
You need two values from the upload step. If you didn't capture them, fetch them now with get_project (note: this one takes the projects/ prefix):
{
"name": "get_project",
"arguments": { "name": "projects/3780309359108792857" }
}
Take the newest entry in screenInstances and read off both id and sourceScreen.
Step 2: Call the MCP tool
{
"name": "create_design_system_from_design_md",
"arguments": {
"projectId": "3780309359108792857",
"selectedScreenInstance": {
"id": "a1b2c3d4e5f6",
"sourceScreen": "projects/3780309359108792857/screens/98b50e2ddc9943efb387052637738f61"
},
"deviceType": "DESKTOP"
}
}
Three different ID formats in one call. This is the single easiest place to get it wrong:
| Field | Format | Example |
|---|---|---|
projectId |
numeric only, no prefix | 3780309359108792857 |
selectedScreenInstance.id |
screen instance id — not the source screen id | a1b2c3d4e5f6 |
selectedScreenInstance.sourceScreen |
full resource path, both prefixes | projects/.../screens/... |
deviceType is optional: MOBILE, DESKTOP, TABLET, or AGNOSTIC. Omit it and Stitch decides.
Step 3: Confirm and capture the asset id
On success you get back a design system with an asset id in assets/NUMERIC_ID form. Store both forms: apply_design_system wants the bare numeric part, generate_screen_from_text's designSystem param wants the full assets/NUMERIC_ID string.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 12d ago First seen · 91 lines · 46 tokens per session scan A 32916f424324
stitch-mcp-create-design-system-from-design-md is a skill published in the GitHub repository gabelul/stitch-kit (45 stars, last pushed 1mo ago), licensed Apache-2.0. It adds 46 tokens to every session and 975 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
design-to-code
Mockup-to-component pipeline using Google Stitch, 21st.dev, and Storybook MCP. Accepts a screenshot, a description, or a URL and produces production-ready React components, checking existing Storybook components before generating anything new. Use when implementing UI from a mockup or screenshot. To call the MCP tool…
kit-montar
Monta um site novo combinando peças já validadas do acervo pessoal do usuário — templates, design systems, componentes, animações, menus, prompts e receitas — em vez de gerar tudo do zero. Acervo vazio: constrói do zero e oferece guardar. Recebe o briefing, decide cor e tipografia antes da identidade, pergunta que…
kit-buscar
Consulta o acervo pessoal de sites do usuário — o que ele já guardou: templates rodáveis, design systems com paleta e tipografia, componentes de UI e de IA, animações de scroll, texto e menu, efeitos WebGL, prompts de página inteira e receitas de composição que já deram certo. Responde "o que eu já tenho pronto pra…
kit-cor
Define a paleta do site como decisão, não como consequência de ter escolhido um design system. Faz as perguntas que determinam cor (o que a marca promete, o que o setor já usa, para onde o olho tem que ir, o que ela NÃO pode parecer), monta a paleta por papel — fundo, tinta, superfície, acento — mede o contraste de…
kit-tipo
Define a tipografia do site como decisão — e, principalmente, resolve se dá para ENTREGAR. Confere licença (site de marca paga por tipo, e acervo montado a partir de site de marca vem cheio de fonte comercial que não se pode servir), propõe substituto livre com critério de desenho, monta o par display + corpo, gera a…
stitch-react-native
Convert Stitch HTML designs into React Native screens, or sync existing native components to updated Stitch designs, using native primitives, StyleSheet rules, and mobile platform checks.