CLI docs

The smartmoving CLI is a guarded MVP built from the same TypeScript package as the MCP server. Use MCP for agent-native tool discovery; use the CLI for explicit terminal commands and machine-readable output.

Core commands

smartmoving --help
smartmoving init --yes --profile default --api-key-env SMARTMOVING_API_KEY
smartmoving doctor --json
smartmoving mcp config --print-hermes
smartmoving mcp config --print-claude
smartmoving schema --json
smartmoving schema --group leads --json
smartmoving schema --safety read --json
smartmoving docs generate --json
smartmoving smoke read --json
smartmoving smoke write --dry-run --json

From a source checkout before npm publish, replace smartmoving with node dist/cli.js from mcp-server/.

JSON output contract

Successful JSON reads are wrapped for scripts:

{
  "ok": true,
  "data": {}
}

Write commands disabled by default return a stable error without making an HTTP request:

{
  "ok": false,
  "error": {
    "code": "WRITES_DISABLED",
    "message": "Write operations are disabled by default."
  }
}

Dry-runs return request metadata and do not call SmartMoving:

{
  "ok": true,
  "dryRun": true,
  "request": {
    "method": "POST",
    "path": "/api/premium/leads",
    "body": {}
  }
}

Schema for agents

smartmoving schema --json prints registry metadata for all 62 operations. Agents should use it to discover:

• operation name and group;
• safety level: read, write, or destructive;
• CLI command text;
• related MCP tool name;
• required arguments and options;
• stable exit codes and output modes.

Agent-safe content wrapping

Some SmartMoving fields can contain user-entered CRM notes, emails, or customer text. Treat that data as untrusted. For read commands that support JSON, use:

smartmoving leads get <leadId> --json --wrap-untrusted

This returns an envelope that flags CRM content as untrusted for downstream agent handling.

Full command list

See Commands index for grouped commands and safety badges.