agentskillexchange DealMachine Sales Intelligence
Use this skill when an agent needs structured US property, owner, people, or company data for sales prospecting, marketing audience research, real estate lead generation, contact enrichment, comparable sales, or a targeted export. It connects to DealMachine through the public dm CLI or the hosted MCP server at https://mcp.dealmachine.com.
The workflow starts by discovering valid filter and field IDs, then counts or estimates the matching audience before running a paid search or export. It limits contact enrichment to the people and fields the user actually requests. For broad or credit-consuming operations, it explains the scope and asks for confirmation first. Results should distinguish counts, previews, enriched records, and completed exports.
Install the skill
Install the canonical upstream skill from the DealMachine repository:
npm exec --yes skills@latest -- add DealMachine/dealmachine-cli --skill dealmachine
For a manual installation, clone the upstream repository and copy skills/dealmachine into the skill directory used by Codex, Claude Code, Cursor, or another compatible agent.
Connect the hosted MCP server
Add the following remote HTTP server to a compatible MCP client:
{
"mcpServers": {
"dealmachine": {
"type": "http",
"url": "https://mcp.dealmachine.com"
}
}
}
The hosted server supports OAuth 2.1. Developer clients that support bearer-token configuration can use a DealMachine API key. Never paste credentials into a prompt or commit them to a repository.
Install the CLI
npm install -g dealmachine
dm login
The CLI provides JSON output for automation and covers property and people search, enrichment, comparable sales, filters, fields, lists, exports, usage, account activity, direct mail, CRM, tasks, and developer utilities. Run dm --help or a command-specific help screen before constructing an unfamiliar request.
Safe operating pattern
- Identify the entity, location, filters, output fields, audience, limit, and desired result type.
- Discover supported filters and fields instead of inventing IDs.
- Run a count or cost estimate before a broad paid operation.
- Start with a small preview unless the user asks for a complete result.
- Confirm the expected scope and credit impact before a material export.
- Report the filters applied, records returned, and credits used when available.
See the upstream repository for current commands, MCP Tool coverage, and authentication details.
Installation
Install or set up from the source-backed instructions:
npm install -g dealmachine