Hospitality Guest Services
Pass
Audited by Gen Agent Trust Hub on Oct 25, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process and respond to untrusted data from external sources, such as guest messages, complaints, and online reviews from platforms like TripAdvisor or Google. This creates an attack surface where malicious instructions could be embedded in guest feedback to influence the agent's behavior.
- Ingestion points: Processes guest emails, feedback forms, and public online reviews (SKILL.md).
- Boundary markers: Absent; there are no instructions provided to use delimiters or safety warnings to separate external guest content from the agent's internal instructions.
- Capability inventory: The skill focuses on generating text-based deliverables and templates; no direct tool access, shell execution, or network operations are defined in this skill.
- Sanitization: No procedures for sanitizing, escaping, or validating guest-provided text are described.
Audit Metadata