hubspot-automation
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests data from HubSpot CRM records, such as contact details, support tickets, and deal notes. If these records contain malicious prompt instructions, they could influence the agent's behavior during automated workflows.
- Ingestion points: Data is fetched via tools such as
HUBSPOT_SEARCH_CONTACTS_BY_CRITERIA,HUBSPOT_GET_TICKET, andHUBSPOT_SEARCH_DEALS(SKILL.md). - Boundary markers: Absent. The instructions do not define delimiters or specific warnings to the agent regarding the untrusted nature of the CRM data.
- Capability inventory: The skill has the ability to search, create, and update multiple CRM objects (Contacts, Companies, Deals, Tickets) and manage connection states.
- Sanitization: Absent. There is no mention of filtering or escaping data retrieved from the CRM.
- [EXTERNAL_DOWNLOADS]: The setup instructions require adding an external MCP server endpoint (
https://rube.app/mcp). This service acts as the gateway for all CRM operations and data transit, representing a third-party dependency.
Audit Metadata