recipe-send-personalized-emails
Pass
Audited by Gen Agent Trust Hub on Sep 19, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from an external source and uses it to trigger further agent actions.
- Ingestion points: The skill retrieves data from a spreadsheet using
gws sheets +readinSKILL.md. - Capability inventory: The skill uses the
gws gmail +sendcommand to send emails based on the ingested data. - Boundary markers: No delimiters or specific instructions are provided to the agent to treat the spreadsheet content strictly as data rather than potential instructions.
- Sanitization: There is no evidence of sanitization or filtering of the spreadsheet content before it is used to populate email fields.
Audit Metadata