Add hubspot-leadgenius partner skill for enterprise CRM sync

[thierryteisseire]

What this PR does

This PR introduces the hubspot-leadgenius skill—a production-grade enterprise data intelligence pipeline. It acts as an autonomous integration bridge, allowing AI Agents to perfectly synchronize heavily-nested AI insights from the LeadGenius Automation API natively into HubSpot CRM.

This serves as a high-fidelity blueprint for enterprise integrations because it demonstrates how Claude can automate custom schema mapping, token bypass execution, and bulk data normalization logic without relying on fragile Zapier-style abstractions.

Key Technical Achievements Demonstrated:

• Global Rate-Limit Bypasses: Actively delegates Administrative Keys (X-Admin-Key) to securely pull thousands of leads without facing the standard HTTP 429 limitations native to traditional scraping arrays.
• Dynamic UI Schema Injection: Bypasses UI limits by hitting HubSpot's crm/v3/properties natively—automatically creating a nested leadgenius_data sidebar group with 40+ custom properties before executing the batch payload.
• Deep JSON Array Decoupling: Employs precise .get('data').get('id') and regex pipeline destructuring to slice deep JSON strings (AiColdEmail, CompanyDescription) into natively readable HubSpot tables devoid of HTML/array artifacts.
• Idempotent Execution: Perfectly maps high-volume data dumps using the crm/v3/objects/contacts/batch/upsert endpoint tied to standard email constraints, neutralizing duplication risks entirely.

Why add this to Partner/Enterprise Skills?

As agentic implementations scale into B2B sales development environments, RevOps engineers need best-in-class scripts for safely normalizing deep qualitative AI data from lead generation tools directly into CRMs. This skill natively mimics your existing salesforce-leadgenius blueprint, giving builders a bulletproof template for interacting with HubSpot's highly rigid schemas.