Do I need a HubSpot API key to use this skill?

No, you don't need a separate HubSpot API key. The skill uses Rube MCP and Composio's HubSpot toolkit with OAuth authentication managed through `RUBE_MANAGE_CONNECTIONS`.

What's the maximum batch size for creating or updating contacts and companies?

The maximum batch size is 100 records per call for both contacts and companies; larger sets require chunking into multiple calls.

Why should I search before creating contacts or companies?

Searching first with `HUBSPOT_SEARCH_CONTACTS_BY_CRITERIA` or `HUBSPOT_SEARCH_COMPANIES` helps you avoid creating duplicate records.

What should I do if ticket searches return zero results?

Fall back to `HUBSPOT_READ_ALL_PROPERTIES_FOR_OBJECT_TYPE` to discover the exact internal property names, since search filters may silently fail if you use display labels instead of internal names.

Skill

Automate HubSpot CRM Operations

Name: HubSpot CRM Automation via Rube MCP
Availability: OnlineOnly
Author: Antigravity

Automates HubSpot CRM contacts, companies, deal pipelines, tickets, and custom properties through Composio's toolkit via Rube MCP.

Get skill

Works with hubspot

Antigravity

Own this? Claim it

Spark score

out of 100

Updated last month

Version 13.1.0

Models

universal

Add to Favorites

Why it matters

Streamline your sales and customer support by automating core HubSpot CRM functions. This asset manages contacts, companies, deals, and tickets, ensuring your data is up-to-date and accessible.

Outcomes

What it gets done

Create and update contacts and companies in HubSpot.

Manage deal pipelines and track progress.

Search and filter support tickets efficiently.

Create and manage custom CRM properties.

Install

Add it to your toolbox

Run in your project directory:

curl -fsSL https://spark.entire.vc/get/ag-hubspot-automation | bash

Overview

HubSpot CRM Automation via Rube MCP

Documents how to automate HubSpot CRM contact, company, deal-pipeline, ticket, and custom-property workflows through Composio's toolkit over Rube MCP, with exact tool sequences and the internal-name/pagination/batch quirks needed to use them correctly. Use when automating HubSpot contacts, companies, deals, tickets, or custom properties via Rube MCP with an active HubSpot connection already in place.

What it does

Automates HubSpot CRM workflows - contact and company management, deal pipeline tracking, ticket search, and custom property creation - through Composio's HubSpot toolkit exposed via Rube MCP. It documents exact tool sequences, required parameters, and the internal-name/pagination/batch-limit quirks needed to drive HubSpot correctly.

When to use - and when NOT to

Use this skill when a task requires creating or updating HubSpot contacts/companies, searching or tracking deals through pipeline stages, searching support tickets, or creating custom CRM properties. It assumes Rube MCP is already connected (RUBE_SEARCH_TOOLS available) and the HubSpot connection is ACTIVE via RUBE_MANAGE_CONNECTIONS - always call RUBE_SEARCH_TOOLS first for current schemas, and run HUBSPOT_GET_ACCOUNT_INFO first since auth failures cascade to every subsequent call.

Inputs and outputs

Setup: add https://rube.app/mcp as an MCP server (no API keys needed), verify RUBE_SEARCH_TOOLS responds, call RUBE_MANAGE_CONNECTIONS with toolkit hubspot, complete the HubSpot OAuth link if not ACTIVE.

Five core workflows: Contacts - search first with HUBSPOT_SEARCH_CONTACTS_BY_CRITERIA to avoid duplicates, then HUBSPOT_CREATE_CONTACT (single) or HUBSPOT_CREATE_CONTACTS (batch, max 100). Companies - HUBSPOT_SEARCH_COMPANIES before HUBSPOT_CREATE_COMPANIES/HUBSPOT_UPDATE_COMPANIES (batch, max 100), storing returned IDs immediately. Deals/pipeline - HUBSPOT_RETRIEVE_ALL_PIPELINES_FOR_SPECIFIED_OBJECT_TYPE to map pipeline/stage IDs first, then HUBSPOT_SEARCH_DEALS filtering on internal property names like pipeline, dealstage, createdate, hubspot_owner_id - results nest under response.data.results with properties often as strings. Tickets - HUBSPOT_SEARCH_TICKETS with filterGroups, falling back to HUBSPOT_READ_ALL_PROPERTIES_FOR_OBJECT_TYPE to discover exact property names when filters silently return zero results. Custom properties - HUBSPOT_READ_ALL_PROPERTIES_FOR_OBJECT_TYPE to check existing fields, then HUBSPOT_CREATE_PROPERTY_FOR_SPECIFIED_OBJECT_TYPE (name/label/type/fieldType/options for enumerations) - property names are immutable once created.

Common patterns: ID resolution (property display name to internal name, pipeline/stage name to ID, owner name to ID all go through dedicated lookup tools); cursor-based pagination via paging.next.after until absent, not page numbers; and batch operations capped at 100 records per call, requiring chunking for larger sets.

Known pitfalls preserved from the source: all search/filter endpoints require internal property names, not display labels; date properties may be epoch-ms or ISO strings depending on endpoint; property names can never be renamed after creation; and always searching before creating contacts/companies to prevent duplicates.

Integrations

Runs entirely through Rube MCP (https://rube.app/mcp) and Composio's HubSpot toolkit - RUBE_SEARCH_TOOLS for schema discovery and RUBE_MANAGE_CONNECTIONS for the HubSpot OAuth connection, with no separate HubSpot API key management required.

Who it's for

Sales/RevOps teams and automation builders who need to drive HubSpot CRM contact, company, deal, ticket, and custom-property workflows through Composio/Rube MCP tools correctly - especially around the internal-name, pagination, and batch-limit quirks that are easy to get wrong from the raw HubSpot API alone.

FAQ

Common questions

Discussion