V0 PlatformSAFE
Overview
From the repository's own README, as read at the audited commit. Badges and raw HTML are left out.
[](https://www.npmjs.com/package/v0-platform-mcp)
Vercel v0 MCP Server for Claude Code - Multi-screen UI prototyping and incremental development through the Model Context Protocol.
Built with Claude Code using Vibe Coding methodology
Quick Start
# 1. Get your v0 API key from https://vercel.com/docs/v0/model-api # 2. Add to Claude Code claude mcp add v0-platform-mcp --env V0_API_KEY=YOUR_KEY -- npx v0-platform-mcp
That's it! Start building prototypes immediately.
Two Workflows
Prototype Workflow (New Projects)
Fast track from idea to implementation in 3 steps:
1. prepare_prototype_context → Parse description into structured context 2. generate_prototype → Generate multi-screen UI prototype 3. handoff_to_claude_dev → Create implementation brief
Use for: MVPs, rapid prototyping, proof-of-concepts
MCP Workflow (Existing Projects)
Incremental development with persistent context in 5 steps:
1. load_project_context → Load existing project state 2. plan_increment → Analyze patterns & plan feature 3. generate_prototype → Generate UI following patterns 4. handoff_to_claude_dev → Create implementation brief 5. update_project_context → Save feature & components
Use for: Production apps, incremental features, pattern consistency
Core Tools
Example: E-commerc
4ed5448a9c83OBSERVED · 2026-10-09Connect
Built from this server's own package name, version and transport as found in its source — not copied from anyone's documentation, so it cannot drift against a page we do not control. Replace the environment placeholders with a token scoped to the least it needs.
claude mcp add v0-platform-mcp --env V0_API_KEY=${V0_API_KEY} -- npx -y [email protected]{
"mcpServers": {
"v0-platform-mcp": {
"command": "npx",
"args": [
"-y",
"[email protected]"
],
"env": {
"V0_API_KEY": "${V0_API_KEY}"
}
}
}
}Exposed tools (7)
6 read · 1 write · 0 destructive.
| Tool | Risk | Description |
|---|---|---|
generate_prototype | read | Generate multi-screen UI prototype using V0 AI based on structured prototype context. Creates React components with TypeScript and Tailwind CSS. Returns prototype with preview URL and component list. This is step 2 in the prototype workflow (after prepare_prototype_context). |
handoff_to_claude_dev | read | Convert a V0 prototype into a structured implementation brief for Claude dev agents. Extracts screens, components, UX patterns, and defines implementation boundaries. This is step 3 in the prototype workflow (after generate_prototype). |
load_project_context | read | Load project context from a file for context-driven incremental development. Reads and validates project state including features, routes, components, and architectural patterns. This is step 1 in the MCP workflow. |
plan_increment | read | Plan how to implement a new feature based on project context. Analyzes existing features, determines generation scope, identifies reusable components, and creates an implementation strategy. This is step 2 in the MCP workflow (after load_project_context). |
prepare_prototype_context | read | Parse natural language product description into structured prototype context. Extracts product name, goal, platform (web/mobile), and infers required screens. This is the first step in the prototype workflow before calling generate_prototype. |
update_project_context | write | Update project context after implementing a feature. Marks features as done, adds new routes, registers new components, and records architectural patterns. This is step 5 in the MCP workflow (after handoff_to_claude_dev and implementation). |
v0_healthcheck | read | Check v0 API configuration and connectivity. Validates API key and endpoint accessibility. |
Trust audit
SAFEgrade B · trust 88/100 Nothing in the source contradicts what it says it does. Grade A is reserved for packages that have also passed the behavioural sandbox.
| Layer | What it checks | Result |
|---|---|---|
| L0 | Provenance & inventory | PASS |
| L1 | Static analysis of the code | PASS |
| L2 | Instruction surface (what it tells the agent) | PASS |
| L3 | Class-specific surface | PASS |
| L4 | Behavioural (sandbox) | SKIPPED |
What the source does
- Filesystem
- none-observed
- Network
- declared (2 observation(s))
- Shell
- none-observed
- Dependencies
- not all pinned
- Secrets in source
- none-found
Findings (7)
import { ContextPreparationService } from '../../src/services/contextPreparationService.js';import { HandoffService } from '../../src/services/handoffService.js';import { V0Tools } from '../../src/mcp/tools.js';import type { PrototypeContext, PrototypeResult } from '../../src/types/index.js';const { config } = await import('../../../src/config/index.js');@modelcontextprotocol/sdk, dotenv, openai, winston, zod, @jest/globals, @types/jest, @types/node
Gates applied: no_behavioural_pass, no_license.
4ed5448a9c83full audit observations/trust-audit/mcp-server/longnguyen0612__v0-platform.json · Report an issue / request a re-scanAudit history
Every audit this server has had. A grade with a past is a grade somebody is still checking.
| Date | Source | Verdict | Grade | Score | Change |
|---|---|---|---|---|---|
| 2026-10-09 | 4ed5448a9c83 | SAFE | B | 88 | first audit |
Questions
What tools does V0 Platform expose?
7 in total: 6 read-only, 1 that write, and 0 that can delete or overwrite. Every one is listed on this page with its risk.
Is V0 Platform safe to connect to an agent?
The audit found nothing in the source that contradicts what it says it does, and graded it B (88/100). Grade A is held back for packages that have also passed a sandboxed behavioural run, which is why a clean server reads B.
What credentials does V0 Platform need?
It reads V0_API_KEY from the environment. Give it a token scoped to the least it needs — an agent that can be talked into calling a tool can be talked into calling it with your credentials.
How does V0 Platform run?
It speaks stdio, so it runs as a local process your client starts. It is published on npm as v0-platform-mcp at 1.3.0.
How current is this page?
The grade is for one exact copy of the source (4ed5448a9c83), read on 2026-10-09. The repository is watched and re-audited when it changes.