The Agent Skills Directory

[SAFE]: No security issues detected. The skill instructions include explicit prohibitions against requesting or repeating sensitive user data such as passwords or credit card numbers.
[SAFE]: The agent is restricted to informational tools and is explicitly directed to offload transactional or sensitive operations to a dedicated support flow, minimizing the risk of unauthorized state changes.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted input from customers and external records, creating a standard injection surface.
Ingestion points: User queries and tool outputs from get_orders and search_policies in SKILL.md.
Boundary markers: None present.
Capability inventory: Limited to read-only data retrieval and display tools (get_orders, search_policies, present_order_status, present_guide).
Sanitization: No automated sanitization, but the skill incorporates manual safety rules for sensitive data.

customer-care