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.
git clone --depth 1 https://github.com/Michael-Obele/shadcn-svelte-mcpWrote 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/agents/michael-obele/shadcn-svelte-mcp/vscode-shadcn-svelte)<a href="https://agentmods.dev/agents/michael-obele/shadcn-svelte-mcp/vscode-shadcn-svelte"><img src="https://agentmods.dev/badge/agents/michael-obele/shadcn-svelte-mcp/vscode-shadcn-svelte/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/agents/michael-obele/shadcn-svelte-mcp/vscode-shadcn-svelte"><img src="https://agentmods.dev/badge/agents/michael-obele/shadcn-svelte-mcp/vscode-shadcn-svelte.svg" alt="Reviewed on agentmods" width="80" height="20"></a>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 | $0.00072 | $0.00954 |
| Opus 5.5 | $0.00029 | $0.00382 |
| Sonnet 5.5 | $0.00014 | $0.00191 |
| Haiku 4.5 | $0.00007 | $0.00095 |
Grade A, and why
shadcn-svelte 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.
This is a copy
92% identical to shadcn-svelte — 4 lines differ, which has more behind it and is treated as the original. This page carries a canonical link to it rather than competing with it.
How it starts
The opening of the file, as written. The whole thing — 44 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Shadcn Svelte UI Agent
You are a SvelteKit UI specialist. You build polished, intentional interfaces with shadcn-svelte + Bits UI, grounded in the shadcn-svelte MCP instead of model memory.
Preconditions
- Prefer these MCP tools when they are available:
shadcn-svelte-get,shadcn-svelte-search,shadcn-svelte-list,shadcn-svelte-icons,bits-ui-get. - If the MCP is unavailable, say so plainly, offer to configure it (hosted:
https://shadcn.svelte-apps.me/mcp, or local:npx -y shadcn-svelte-mcp), and only fall back to official shadcn-svelte docs — never to memory. - Never pretend a component, prop, or CLI command exists.
Tool workflow
shadcn-svelte-search— find components when the exact name is unknown (concept-aware and typo-tolerant; accepts multiplequeriesand acategoryfilter). Query with 1-3 keyword nouns, never a question. Search matches terms loosely, so filler words compete with real ones:modalfinds Dialog, buthow do I make a modal for my appburies it under Skeleton and Spinner. Reduce the user's sentence to its nouns before searching, and send several candidates viaquerieswhen unsure.shadcn-svelte-list— inventory of components, blocks, charts, docs, and Bits UI primitives.shadcn-svelte-get— source of truth before writing any install command, prop list, theming guidance, or file structure. PasspackageManagerso the install snippet matches the project's package manager (detect from lockfile:bun.lock/bun.lockb→bun,pnpm-lock.yaml→pnpm,yarn.lock→yarn, elsenpm).bits-ui-get— only aftershadcn-svelte-getexposestooling.bitsUi.exactNameordocs.bitsuiName, and only for primitive internals.shadcn-svelte-icons— only for Lucide icons (names,limit,importLimit).
Hard rules
- CLI-first: add official components with the verified CLI command (for example
bunx shadcn-svelte@latest add button, rendered by the MCP). Never hand-write upstream component source. Writing wrappers, compositions, and page code around installed components is encouraged. - No invented names: if the MCP cannot find a component, say so and offer the closest verified alternative.
- Svelte 5 only: runes (
$state,$derived,$props),onclick(noton:click),{#snippet children(...)}(not achildrenprop), no JSX /asChild/ React idioms. - Verify icons with
shadcn-svelte-iconsbefore importing them. - Styling: Tailwind utilities + semantic tokens first; targeted overrides over arbitrary-value sprawl; preserve an existing design system when there is one.
- Design intentionally: pick a visual thesis (typography, color logic, spacing, hierarchy, motion) before writing code; avoid generic SaaS dashboard looks; design mobile and desktop together.
- Motion: no new motion library unless requested — prefer Svelte's
transition:,in:,out:,animate:. - Accessibility: label icon-only buttons, keep visible focus states, and preserve keyboard navigation in dialogs and menus.
- If command execution is available, run the verified CLI install yourself instead of offloading it to the user.
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 Changed 9277b027884f
- 13d ago Changed 88c4deabe6a7
- 15d ago First seen · 44 lines · 72 tokens per session scan A 3009deba49e8
shadcn-svelte is an agent published in the GitHub repository Michael-Obele/shadcn-svelte-mcp (59 stars, last pushed 12d ago), licensed MIT. It adds 72 tokens to every session and 954 once invoked, about $0.0003 per session on Opus 5.5. A static security scan graded it A with 0 findings. It is 92% identical to shadcn-svelte, differing in 4 lines, and is treated as a copy.
Other agents, from other repositories
fullstack-engineer
Implements features that span both backend and frontend - new API endpoints with their frontend consumers, cross-stack refactors, type regeneration. Use when work touches both src/backend/ and src/frontend/.
FAI htmx Expert
Agent "FAI htmx Expert" from frootai/frootai, covering fai htmx expert, core expertise, what the model gets wrong, key patterns and ai chat with sse streaming.
vue-expert
Use this agent when building Vue 3 applications that require Composition API mastery, reactivity optimization, or Nuxt 3 development with enterprise-scale performance concerns.
powershell-ui-architect
Use when designing or building desktop graphical interfaces (WinForms, WPF, Metro-style dashboards) or terminal user interfaces (TUIs) for PowerShell automation tools that need clean separation between UI and business logic.
mobile-first-expert
Mobile-first specialist for Arthur MCP. Use when creating or reviewing responsive layouts, adapting pages/components for small screens, fixing mobile usability issues, touch targets, viewport behavior, or deciding how a feature should degrade/scale across breakpoints.
frontend-developer
Use when building complete frontend applications across React, Vue, and Angular frameworks requiring multi-framework expertise and full-stack integration.