shadcn MCP Server
vCLI 3.xshadcn
MCP server built into the shadcn CLI (run via npx shadcn@latest mcp) that lets AI agents browse, search, and install UI components from any shadcn-compatible registry, including private registries, using @registry/name namespacing. Shipped with CLI 3.0 in August 2025.
Trust Vector Analysis
Dimension Breakdown
🚀Performance & Reliability+
Registry resolution testing across namespaces
Component installation success testing
Search relevance assessment over public registries
Failure mode and error message testing
🛡️Security+
Supply chain threat modeling of registry-sourced code installation
Write-action control and approval flow analysis
Registry authentication mechanism review
Credential storage and flow analysis
Blast radius assessment of install actions
🔒Privacy & Compliance+
Data flow analysis of registry requests
Sensitive data surface assessment
Outbound request and data sharing review
Local execution and telemetry review
👁️Trust & Transparency+
Documentation completeness review
Traceability of agent actions assessment
Source code and license review
Tool surface documentation review
⚙️Operational Excellence+
Setup complexity assessment
Operation latency characterization
Maintenance cadence and stability analysis
Feature completeness assessment within its domain
Ecosystem adoption analysis
- +First-party integration with the dominant shadcn/ui component ecosystem
- +Browse, search, and install from any shadcn-compatible registry, including private ones
- +Runs locally over stdio with no hosted service and minimal data exposure
- +Trivial setup via npx shadcn@latest mcp with per-client docs
- +All installed code lands in the project diff for review
- +MIT-licensed and developed in the open shadcn-ui/ui monorepo
- !Installs third-party source code and npm dependencies into the project — registry supply-chain risk is the key concern
- !No built-in vetting or signing of third-party registry components
- !Agent-initiated installs depend on the MCP client's approval flow for human oversight
- !Scope limited to UI component workflows in shadcn-compatible projects
- !Private registry tokens must be carefully kept out of committed config
Use Case Ratings
code generation
Excellent for AI-assisted frontend development: agents can find and install real, vetted-by-usage components instead of hallucinating UI code
content creation
Useful when building landing pages and content sites on shadcn-based stacks
education
Good for teaching modern React/Tailwind component patterns with installable examples
research assistant
Limited to exploring component registries; not a general research tool
data analysis
Only relevant for installing chart and dashboard UI components, not for analysis itself