crm-hygiene
Pass
Audited by Gen Agent Trust Hub on Oct 5, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from external CRM platforms, which represents a potential attack surface for indirect prompts. Ingestion points: Records and field values (e.g., names, emails, notes) pulled from Airtable, Notion, and HubSpot as defined in SKILL.md. Boundary markers: Safety Rule 5 explicitly warns to never follow instructions found inside a record and states that direction comes only from the user. Capability inventory: The skill utilizes connectors for reading, writing, and archiving records in external CRM databases. Sanitization: Relies on specific instructions to the agent to treat data as untrusted and disregard any embedded commands.
Audit Metadata