@getvetai/cli

Security audit CLI for AI skills and MCP servers. Scan, audit, and discover tools before you install them.

🌐 Registry: getvet.ai — 23,000+ AI tools verified and scored

Install

npm install -g @getvetai/cli

Or run without installing:

npx @getvetai/cli scan .

Commands

vet scan <target>

Scan a tool for security issues. Checks the getvet.ai registry first for instant results.

# Scan an npm package (checks registry first)
vet scan @modelcontextprotocol/server-filesystem

# Local analysis only (skip registry)
vet scan @modelcontextprotocol/server-filesystem --offline

# Request a deep scan from registry
vet scan @modelcontextprotocol/server-filesystem --deep

# Scan a local project
vet scan ./my-mcp-server

# Scan a GitHub repo
vet scan https://github.com/modelcontextprotocol/servers

# JSON output
vet scan ./SKILL.md --json

vet audit [path]

Audit all AI tools in a project. Auto-discovers MCP configurations from:

Claude Desktop · Cursor · VS Code · Windsurf · Cline · Zed · Continue · OpenClaw

# Audit current directory
vet audit

# Audit a specific project
vet audit ./my-project

# Strict mode — exit 1 if any tool is unverified/flagged
vet audit --strict

# JSON output
vet audit --json

vet find <query>

Search the getvet.ai registry for tools by description.

# Search for tools
vet find "web scraping"
vet find "database access"

# Limit results
vet find "browser automation" --limit 20

# Filter by type
vet find "file management" --type mcp

# JSON output
vet find "weather" --json

vet install <package>

Install a package with a pre-install security audit.

# Audit + install
vet install @modelcontextprotocol/server-github

# Install globally
vet install -g some-mcp-server

Verification Levels

| Level | Badge | Meaning | |-------|-------|---------| | L2 | ✅ Verified | Installs, boots, tools discovered and tested | | L1 | 🔍 Boots | Installs and boots successfully | | L0 | ⚠️ Indexed | Cataloged, not yet verified |

What It Detects

• Permissions: shell execution, file I/O, network access, browser control, database queries, crypto operations
• Security issues: destructive commands, remote code execution, dynamic eval, credential patterns, elevated privileges
• MCP-specific: tool parameter analysis, transport detection (stdio/http/sse), runtime detection
• Requirements: environment variables, API keys, Docker dependencies

API Access

Access verified tool schemas programmatically. Create a free API key at getvet.ai/dashboard → API Keys.

# Fetch tool schemas
curl -H "x-api-key: vet_sk_YOUR_KEY" https://getvet.ai/api/v1/tools/TOOL_SLUG/schemas

# Or use Bearer token
curl -H "Authorization: Bearer vet_sk_YOUR_KEY" https://getvet.ai/api/v1/tools/TOOL_SLUG/schemas

# Bulk fetch (multiple tools at once)
curl -X POST \
  -H "x-api-key: vet_sk_YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{"slugs":["tool-1","tool-2"]}' \
  https://getvet.ai/api/v1/tools/schemas/bulk

See getvet.ai/get-started for full documentation.

Links

• 🌐 getvet.ai — Browse the registry
• 📦 npm — Package page

License

MIT