Install to first finding in one command. Node 18 or newer.
npx mcpvetter scannpx mcpvetter scan # discover configs + static scan npx mcpvetter scan --live # also start and inspect servers (asks first, per server) npx mcpvetter scan --json out.json # machine-readable report npx mcpvetter scan --fail-on high # exit 1 if any finding is high or worse (CI) npx mcpvetter list # inventory only npx mcpvetter explain MCP-SECRET-001
Exit codes: 0 no findings at or above the threshold, 1 findings at or above it, 2 the scan could not complete.
The local server runs every check on your machine. Without a key you get the free scanner tools; add an API key from your account to unlock Pro tools.
claude mcp add mcpvetter -e MCPVETTER_API_KEY=mcpv_live_YOUR_KEY -- npx -y mcpvetter serve{
"mcpServers": {
"mcpvetter": {
"command": "npx",
"args": [
"-y",
"mcpvetter",
"serve"
],
"env": {
"MCPVETTER_API_KEY": "mcpv_live_YOUR_KEY"
}
}
}
}
{
"servers": {
"mcpvetter": {
"type": "stdio",
"command": "npx",
"args": [
"-y",
"mcpvetter",
"serve"
],
"env": {
"MCPVETTER_API_KEY": "mcpv_live_YOUR_KEY"
}
}
}
}
https://mcpvetter.com/mcpStreamable HTTP with OAuth: your client opens a sign-in page the first time. Clients that support headers can send Authorization: Bearer mcpv_live_YOUR_KEY instead. Code sent to the remote endpoint is analysed in memory and not stored.
claude mcp add --transport http mcpvetter-remote https://mcpvetter.com/mcp| Tool | Plan | What it does |
|---|---|---|
check_mcp_setup | Free | MCP posture summary with fixes |
explain_finding | Free | Plain-English explanation of a finding ID |
review_code | Pro (5/day free) | Findings for a snippet, each linked to a pattern and lab |
secure_pattern | Pro (2 free) | Secure example in your stack, anti-pattern, reviewer checklist |
start_lab / check_lab | Pro (2 free) | 5-minute lab and automatic verification |
team_posture | Team | Aggregated counts for your team |