suppress-global-unsubscribes
Pass
Audited by Gen Agent Trust Hub on Sep 29, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests and processes contact properties like names, emails, and lifecycle stages from the HubSpot CRM Search API, which are considered untrusted external data.\n
- Ingestion points: Contact data is retrieved from the HubSpot API in
scripts/before.pyandscripts/after.pyand then printed to the console or saved to files.\n - Boundary markers: The instructions do not include delimiters or specific guidance for the agent to treat the retrieved contact data as untrusted content, potentially allowing embedded instructions in those fields to influence agent behavior.\n
- Capability inventory: The skill utilizes network capabilities via the
requestslibrary to communicate withapi.hubapi.comand has file system write access to thedata/audit-logs/directory.\n - Sanitization: The script extracts and uses contact properties directly from the API response without performing validation, filtering, or escaping.
Audit Metadata