本文へ移動
cccskills
無料GitHub で公開

hubspot-automation

Automate HubSpot CRM operations (contacts, companies, deals, tickets, properties) via Rube MCP using Composio integration.

インストール方法を見る

含まれるファイル(1)

  • SKILL.md9.2 KB

SKILL.md(原文)

インストールする前に、エージェントに与えられる指示の中身を確認できます。

HubSpot CRM Automation via Rube MCP

Automate HubSpot CRM workflows including contact/company management, deal pipeline tracking, ticket search, and custom property creation through Composio's HubSpot toolkit.

Toolkit docs: composio.dev/toolkits/hubspot

Prerequisites

  • Rube MCP must be connected (RUBE_SEARCH_TOOLS available)
  • Active HubSpot connection via RUBE_MANAGE_CONNECTIONS with toolkit hubspot
  • Always call RUBE_SEARCH_TOOLS first to get current tool schemas

Setup

Get Rube MCP: Add https://rube.app/mcp as an MCP server in your client configuration. No API keys needed — just add the endpoint and it works.

  1. Verify Rube MCP is available by confirming RUBE_SEARCH_TOOLS responds
  2. Call RUBE_MANAGE_CONNECTIONS with toolkit hubspot
  3. If connection is not ACTIVE, follow the returned auth link to complete HubSpot OAuth
  4. Confirm connection status shows ACTIVE before running any workflows

Core Workflows

1. Create and Manage Contacts

When to use: User wants to create new contacts or update existing ones in HubSpot CRM

Tool sequence:

  1. HUBSPOT_GET_ACCOUNT_INFO - Verify connection and permissions (Prerequisite)
  2. HUBSPOT_SEARCH_CONTACTS_BY_CRITERIA - Search for existing contacts to avoid duplicates (Prerequisite)
  3. HUBSPOT_READ_A_CRM_PROPERTY_BY_NAME - Check property metadata for constrained values (Optional)
  4. HUBSPOT_CREATE_CONTACT - Create a single contact (Required)
  5. HUBSPOT_CREATE_CONTACTS - Batch create contacts up to 100 (Alternative)

Key parameters:

  • HUBSPOT_CREATE_CONTACT: properties object with email, firstname, lastname, phone, company
  • HUBSPOT_CREATE_CONTACTS: inputs array of {properties} objects, max 100 per batch
  • HUBSPOT_SEARCH_CONTACTS_BY_CRITERIA: filterGroups array with {filters: [{propertyName, operator, value}]}, properties array of fields to return

Pitfalls:

  • Max 100 records per batch; chunk larger imports
  • 400 'Property values were not valid' if using incorrect property names or enum values
  • Always search before creating to avoid duplicates
  • Auth errors from GET_ACCOUNT_INFO mean all subsequent calls will fail

2. Manage Companies

When to use: User wants to create, search, or update company records

Tool sequence:

  1. HUBSPOT_SEARCH_COMPANIES - Search existing companies (Prerequisite)
  2. HUBSPOT_CREATE_COMPANIES - Batch create companies, max 100 (Required)
  3. HUBSPOT_UPDATE_COMPANIES - Batch update existing companies (Alternative)
  4. HUBSPOT_GET_COMPANY - Get single company details (Optional)
  5. HUBSPOT_BATCH_READ_COMPANIES_BY_PROPERTIES - Bulk read companies by property values (Optional)

Key parameters:

  • HUBSPOT_CREATE_COMPANIES: inputs array of {properties} objects, max 100
  • HUBSPOT_SEARCH_COMPANIES: filterGroups, properties, sorts, limit, after (pagination cursor)

Pitfalls:

  • Max 100 per batch; chunk larger sets
  • Store returned IDs immediately for downstream operations
  • Property values must match exact internal names, not display labels

3. Manage Deals and Pipeline

When to use: User wants to search deals, view pipeline stages, or track deal progress

Tool sequence:

  1. HUBSPOT_RETRIEVE_ALL_PIPELINES_FOR_SPECIFIED_OBJECT_TYPE - Map pipeline and stage IDs/names (Prerequisite)
  2. HUBSPOT_SEARCH_DEALS - Search deals with filters (Required)
  3. HUBSPOT_RETRIEVE_PIPELINE_STAGES - Get stage details for one pipeline (Optional)
  4. HUBSPOT_RETRIEVE_OWNERS - Get owner/rep details (Optional)
  5. HUBSPOT_GET_DEAL - Get single deal details (Optional)
  6. HUBSPOT_LIST_DEALS - List all deals without filters (Fallback)

Key parameters:

  • HUBSPOT_SEARCH_DEALS: filterGroups with filters on pipeline, dealstage, createdate, closedate, hubspot_owner_id; properties, sorts, limit, after
  • HUBSPOT_RETRIEVE_ALL_PIPELINES_FOR_SPECIFIED_OBJECT_TYPE: objectType set to 'deals'

Pitfalls:

  • Results nested under response.data.results; properties are often strings (amounts, dates)
  • Stage IDs may be readable strings or opaque numeric IDs; use label field for display
  • Filters must use internal property names (pipeline, dealstage, createdate), not display names
  • Paginate via paging.next.after until absent

4. Search and Filter Tickets

When to use: User wants to find support tickets by status, date, or criteria

Tool sequence:

  1. HUBSPOT_SEARCH_TICKETS - Search with filterGroups (Required)
  2. HUBSPOT_READ_ALL_PROPERTIES_FOR_OBJECT_TYPE - Discover available property names (Fallback)
  3. HUBSPOT_GET_TICKET - Get single ticket details (Optional)
  4. HUBSPOT_GET_TICKETS - Bulk fetch tickets by IDs (Optional)

Key parameters:

  • HUBSPOT_SEARCH_TICKETS: filterGroups, properties (only listed fields are returned), sorts, limit, after

Pitfalls:

  • Incorrect propertyName/operator returns zero results without errors
  • Date filtering may require epoch-ms bounds; mixing formats causes mismatches
  • Only fields in the properties array are returned; missing ones break downstream logic
  • Use READ_ALL_PROPERTIES to discover exact internal property names

5. Create and Manage Custom Properties

When to use: User wants to add custom fields to CRM objects

Tool sequence:

  1. HUBSPOT_READ_ALL_PROPERTIES_FOR_OBJECT_TYPE - List existing properties (Prerequisite)
  2. HUBSPOT_READ_PROPERTY_GROUPS_FOR_OBJECT_TYPE - List property groups (Optional)
  3. HUBSPOT_CREATE_PROPERTY_FOR_SPECIFIED_OBJECT_TYPE - Create a single property (Required)
  4. HUBSPOT_CREATE_BATCH_OF_PROPERTIES - Batch create properties (Alternative)
  5. HUBSPOT_UPDATE_SPECIFIC_CRM_PROPERTY - Update existing property definition (Optional)

Key parameters:

  • HUBSPOT_CREATE_PROPERTY_FOR_SPECIFIED_OBJECT_TYPE: objectType, name, label, type (string/number/date/enumeration), fieldType, groupName, options (for enumerations)

Pitfalls:

  • Property names are immutable after creation; choose carefully
  • Enumeration options must be pre-defined with value and label
  • Group must exist before assigning properties to it

Common Patterns

ID Resolution

  • Property display name → internal name: Use HUBSPOT_READ_ALL_PROPERTIES_FOR_OBJECT_TYPE
  • Pipeline name → pipeline ID: Use HUBSPOT_RETRIEVE_ALL_PIPELINES_FOR_SPECIFIED_OBJECT_TYPE
  • Stage name → stage ID: Extract from pipeline stages response
  • Owner name → owner ID: Use HUBSPOT_RETRIEVE_OWNERS

Pagination

  • Search endpoints use cursor-based pagination
  • Follow paging.next.after until absent
  • Typical limit: 100 records per page
  • Pass after value from previous response to get next page

Batch Operations

  • Most create/update endpoints support batching with max 100 records per call
  • For larger datasets, chunk into groups of 100
  • Store returned IDs from each batch before proceeding
  • Use batch endpoints (CREATE_CONTACTS, CREATE_COMPANIES, UPDATE_COMPANIES) instead of single-record endpoints for efficiency

Known Pitfalls

  • Property names: All search/filter endpoints use internal property names, NOT display labels. Always call READ_ALL_PROPERTIES_FOR_OBJECT_TYPE to discover correct names
  • Batch limits: Max 100 records per batch operation. Larger sets must be chunked
  • Response structure: Search results are nested under response.data.results with properties as string values
  • Date formats: Date properties may be epoch-ms or ISO strings depending on endpoint. Parse defensively
  • Immutable names: Property names cannot be changed after creation. Plan naming conventions carefully
  • Cursor pagination: Use paging.next.after cursor, not page numbers. Continue until after is absent
  • Duplicate prevention: Always search before creating contacts/companies to avoid duplicates
  • Auth verification: Run HUBSPOT_GET_ACCOUNT_INFO first; auth failures cascade to all subsequent calls

Quick Reference

TaskTool SlugKey Params
Create contactHUBSPOT_CREATE_CONTACTproperties: {email, firstname, lastname}
Batch create contactsHUBSPOT_CREATE_CONTACTSinputs: [{properties}] (max 100)
Search contactsHUBSPOT_SEARCH_CONTACTS_BY_CRITERIAfilterGroups, properties, limit, after
Create companiesHUBSPOT_CREATE_COMPANIESinputs: [{properties}] (max 100)
Search companiesHUBSPOT_SEARCH_COMPANIESfilterGroups, properties, after
Search dealsHUBSPOT_SEARCH_DEALSfilterGroups, properties, after
Get pipelinesHUBSPOT_RETRIEVE_ALL_PIPELINES_FOR_SPECIFIED_OBJECT_TYPEobjectType: 'deals'
Search ticketsHUBSPOT_SEARCH_TICKETSfilterGroups, properties, after
List propertiesHUBSPOT_READ_ALL_PROPERTIES_FOR_OBJECT_TYPEobjectType
Create propertyHUBSPOT_CREATE_PROPERTY_FOR_SPECIFIED_OBJECT_TYPEobjectType, name, label, type, fieldType
Get ownersHUBSPOT_RETRIEVE_OWNERSNone
Verify connectionHUBSPOT_GET_ACCOUNT_INFONone

Powered by Composio

レビュー

まだレビューはありません。使ってみた感想をお寄せください。

同じリポジトリのスキル

概要と使いどころ

Automate ActiveCampaign tasks via Rube MCP (Composio): manage contacts, tags, list subscriptions, automation enrollment, and tasks. Always search tools first for current schemas.

日本語の概要は準備中です。原文の説明を表示しています。

davepoon/buildwithclaude3,6322026年10月12日 更新

Analytics your AI agent can actually use. Track, analyze, run A/B experiments, and optimize across all your projects via CLI. Includes a growth playbook so your agent knows HOW to grow, not just what to track.

日本語の概要は準備中です。原文の説明を表示しています。

davepoon/buildwithclaude3,6322026年10月12日 更新

Teaches when to recall from long-term memory before acting and when to save durable decisions, corrections and failures afterwards. Use when a memory tool or MCP memory server is connected but the agent is not using it consistently, when the user complains that the assistant forgets preferences, conventions or past decisions between sessions, or when setting up persistent memory for a project. Works with any memory backend: a folder of Markdown files, a local MCP server, or a managed service.

日本語の概要は準備中です。原文の説明を表示しています。

davepoon/buildwithclaude3,6322026年10月12日 更新

Audit local AI coding-agent sessions with agenttrace for cost, tokens, tool failures, latency, anomalies, health, diffs, and CI gates.

日本語の概要は準備中です。原文の説明を表示しています。

davepoon/buildwithclaude3,6322026年10月12日 更新

Audit whether a website can be found, crawled, and cited by AI answer engines such as ChatGPT Search, Perplexity, Google AI Overviews, and Microsoft Copilot. Use when someone asks why their brand is missing from AI answers, whether AI crawlers can read their site, how to get cited by ChatGPT or Perplexity, or asks for a GEO or AEO (generative / answer engine optimization) review. Produces a citation baseline across buyer-intent prompts, a crawler-access check, a citability review of named pages, and a ranked fix list. Not for keyword rank tracking, paid search, or pages behind a login.

日本語の概要は準備中です。原文の説明を表示しています。

davepoon/buildwithclaude3,6322026年10月12日 更新

Automate Airtable tasks via Rube MCP (Composio): records, bases, tables, fields, views. Always search tools first for current schemas.

日本語の概要は準備中です。原文の説明を表示しています。

davepoon/buildwithclaude3,6322026年10月12日 更新

davepoon のスキルをすべて見る

このスキルの問題を報告する