handle-complaint
Pass
Audited by Gen Agent Trust Hub on May 13, 2026
Risk Level: SAFE
Full Analysis
- Human-in-the-Loop Controls: The skill includes mandatory approval gates for sensitive operations. It explicitly instructs the agent to never send responses, issue refunds, or close tickets without direct owner confirmation.
- External Service Integration: The skill references integration with well-known services (Gmail, HubSpot, PayPal) to gather customer history and transaction status. These interactions are consistent with the skill's functional requirements.
- Input Processing Safety: While the skill handles external data from customer complaints (which is a potential surface for indirect prompt injection), the risk is addressed by the requirement that all drafts and suggestions be presented to the user for review before any action is taken.
- Tool Usage Context: The skill utilizes standard tools like Bash and WebFetch. Based on the instructions, these appear to be intended for local data parsing and API communication with the specified customer service platforms.
Audit Metadata