crm-sales-momentum
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from HubSpot (deal names, source properties, and owner details) and utilizes potentially powerful tools like
BashandWrite. While this presents a theoretical surface for indirect prompt injection if deal metadata contained malicious instructions, the skill's logic is focused on aggregate mathematical analysis (averages, standard deviations, and conversion percentages) and structured reporting, which significantly mitigates this risk. - Ingestion points: Data is ingested from HubSpot deal, contact, and company objects via
hubspot__search_crm_objectsinSKILL.md. - Boundary markers: Not explicitly defined in the instructions.
- Capability inventory: Access to
Bash,Read,Write, andmcp__cogny__create_findingtools. - Sanitization: Not explicitly defined, though the focus on numeric analysis limits exploitation.
- [EXTERNAL_DOWNLOADS]: The skill interacts with HubSpot CRM APIs via vendor-provided tools. These operations are essential to the primary function of CRM analysis and originate from a trusted vendor environment.
Audit Metadata