persona-customer-support
Pass
Audited by Gen Agent Trust Hub on Oct 18, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill triages and processes untrusted customer emails from Gmail, creating a surface for indirect prompt injection where malicious instructions embedded in an email could influence agent behavior.
- Ingestion points: Gmail inbox queried via
gws gmail +triage(SKILL.md). - Boundary markers: Absent; no instructions provided to delimit or ignore embedded instructions in external email content.
- Capability inventory: Read/write access to Gmail, Google Sheets, Google Chat, and Google Calendar via required skills (SKILL.md).
- Sanitization: Absent; no validation or escaping of email content before processing into tasks or responses.
Audit Metadata