Before your sales team enters a new market, they need to know who's in it. In the old world, finding that out typically meant hours of manual research — searching Google Maps, copying business names into spreadsheets, chasing down contact details one by one.

In the new world, you’d probably start by asking Cursor to handle this for you. However, Cursor’s web search isn’t quite built for this use case. Cursor has a built-in, general-purpose search via WebSearch tool:

• Returns standard search results (titles, snippets, URLs)
• Single search mode, no domain/time filtering
• No structured data extraction

So, to get this info, you’d probably need to set up a web scraper, which breaks whenever sites change and requires hours of configuration.

What if instead, you just told Cursor:

Find all freight logistics companies in Dallas, TX

And got back a structured list of every relevant business; names, addresses, ratings, and contact details pulled live from Google Maps. Ready to work with, immediately.

That's one example of what Nimble Skills makes possible. Whether you're mapping competitive landscapes, building prospect lists, or tracking market activity at scale, the same pattern applies: Nimble's entire web data infrastructure, accessible through Cursor, with no scraping setup and no manual data collection.

To show how it works in practice, we'll walk through a local business discovery use case — mapping businesses across multiple cities using Nimble's pre-built Google Maps agent.

Nimble Skills: Turn Your AI Assistant into a Web Data Expert

Nimble Skills are plug-and-play extensions that connect Cursor directly to live web data. Install them once and Cursor gains the ability to run pre-built extraction agents, search the web in real time, pull structured content from any URL, and build repeatable data pipelines driven entirely by natural language instructions.

There are two skills:

• nimble-web-expert is your interface to the live web. It routes your request to the right tool automatically whether that's a pre-built agent for a known platform, a direct URL extraction, a web search, or a full site crawl. Bot protection, JavaScript rendering, and output parsing are all handled behind the scenes.
• nimble-agent-builder is how you turn a one-off data pull into a production workflow. When you need a site that isn't covered by a pre-built agent, it generates a custom one from a plain language description, built in real-time using AI by Nimble, and immediately available to run at scale.

Both skills work independently and each has its own range of use cases. How you combine them depends on what you're building.

Getting Started

You'll need a Nimble API key from Account Settings. Then follow two steps:

Step 1: Add the Nimble MCP server to Cursor

Open Cursor Settings > Tools & MCP and click Add Custom MCP. Enter the following:

Name: nimble-mcp-server
Type: streamableHttp
URL: https://mcp.nimbleway.com/mcp
Headers: Authorization: Bearer YOUR_API_KEY

Click Install. The nimble-mcp-server will appear under Plugin MCP Servers with 18 tools enabled.

Step 2: Install skills and rules

npx skills add Nimbleway/agent-skills -a cursor

Cursor now has access to both skills.

From First Pull to Full Pipeline

A reliable pattern for teams building data workflows is to validate before you scale.

nimble-web-expert makes validation effortless — run the agent once, inspect the output, confirm the data is accurate and complete. At this stage you're committing nothing. If the results don't look right, you've lost two minutes. If they do, you have everything you need to build around it.

That's when nimble-agent-builder earns its place. Take the validated pull, define the inputs you want to parameterize — city, business type, region — and let it set up a pipeline that runs the same extraction reliably across any number of inputs on whatever schedule your team needs.

This is one way to use the skills together, not the only way. But as a path from idea to production data feed, it's hard to beat for speed.

Mapping Local Business Markets with Google Maps

Here's how this plays out for a team doing local business discovery across multiple cities.

Step 1: The first pull

Ask Cursor to find businesses matching a specific profile in a target city:

Find all freight logistics companies in Dallas, TX

Cursor uses nimble-web-expert to run Nimble's pre-built google_maps_search agent:

nimble agent google_maps_search --params '{"query": "freight logistics companies Dallas TX"}'

You get back structured records pulled live from Google Maps — business details alongside review data for each result:

[
  {
    "place_id": "ChIJsw8_vmOf044RoKPFg7iQFJU",
    "entity_type": "Review",
    "address": "2845 Irving Blvd, Dallas, TX 75247",
    "business_category": ["Freight Forwarding Service", "Transportation"],
    "rating": 5,
    "description": "Reliable partner for cross-border shipments. On-time delivery every time.",
    "username": "Mike T.",
    "user_review_count": 47,
    "relative_time": "2 months ago"
  },
  {
    "place_id": "ChIJsw8_vmOf044RoKPFg7iQFJU",
    "entity_type": "Review",
    "address": "2845 Irving Blvd, Dallas, TX 75247",
    "business_category": ["Freight Forwarding Service", "Transportation"],
    "rating": 4,
    "description": "Good pricing and responsive account managers. Tracking could be better.",
    "username": "Sarah L.",
    "user_review_count": 12,
    "relative_time": "4 months ago"
  }
]

Review the output. Are the business categories accurate? Are the fields populated consistently? If yes, you've confirmed the data source is solid in under a minute - no custom parsing required.

Step 2: Scale across markets

Once the data checks out, tell Cursor you want this running across your full list of target cities:

Run this for freight logistics companies across Dallas, Houston, Phoenix, and Atlanta.
Combine the results into a single prospect list.

That's the signal for Cursor to hand off to nimble-agent-builder, which sets up the pipeline using nimble agent run across each city:

nimble agent run --agent google_maps_search --params '{"query": "freight logistics companies Dallas TX"}'
nimble agent run --agent google_maps_search --params '{"query": "freight logistics companies Houston TX"}'
nimble agent run --agent google_maps_search --params '{"query": "freight logistics companies Phoenix AZ"}'
nimble agent run --agent google_maps_search --params '{"query": "freight logistics companies Atlanta GA"}'

The aggregated output is a structured prospect list across four markets, ready to pipe into your CRM, feed into a report, or hand straight to the sales team.

Taking It Further

Local business discovery is one application. The same two-skill approach works across any domain where you need live, structured data from the web at scale.

Nimble's pre-built agent library covers the platforms businesses rely on most. A few examples:

E-commerce intelligence: Monitor product listings, pricing changes, and availability across major retail platforms:

nimble agent run --agent amazon_pdp --params '{"asin": "B08N5WRWNW"}'
nimble agent run --agent walmart_pdp --params '{"product_id": "123456"}'

Property market tracking: Pull active listings from Zillow by zip code and listing type — sales, rentals, or recently sold:

nimble agent run --agent zillow_plp --params '{"zip_code": "78701", "listing_type": "sales"}'

Search visibility monitoring: Track how your brand or products appear in Google Search results and AI Overviews:

nimble agent run --agent google_search --params '{"query": "best freight software 2025"}'
nimble agent run --agent google_search_aio --params '{"query": "best freight software 2025"}'

Any site, not just prebuilt agents. When there's no pre-built agent for your target, nimble-agent-builder handles the build. Describe what you need in plain language — Nimble generates the agent, Cursor monitors the process and publishes it to your library once it's ready. From that point it runs exactly like any other agent — parameterized, reliable, and ready to scale.

The pattern is always the same. The data source is up to you.

Get started at nimbleway.com.