freshsales
Pass
Audited by Gen Agent Trust Hub on Oct 4, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill documents methods for ingesting external data from Freshworks CRM, which creates a standard integration surface where malicious content in the source CRM could potentially influence agent behavior.
- Ingestion points: Data is ingested via
apideck.crm.contacts.listand the Apideck Proxy API as described inSKILL.md. - Boundary markers: The instructions do not specify explicit delimiters or warnings for the agent to treat the external CRM data as untrusted content.
- Capability inventory: The skill enables the agent to read, search, and write CRM records (contacts, leads, opportunities) and execute arbitrary downstream API calls via the
proxyendpoint. - Sanitization: No specific sanitization or validation logic is defined within the skill for handling the content retrieved from the external service.
Audit Metadata