review-responder
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted content (customer reviews) which creates an attack surface for indirect prompt injection.
- Ingestion points: The 'Review text' placeholder in SKILL.md is populated with external, potentially attacker-controlled content from platforms like Google or Yelp.
- Boundary markers: The prompt template lacks explicit delimiters or instructions telling the agent to ignore command-like content embedded within the review text.
- Capability inventory: The skill's capabilities are limited to text generation; it does not have access to sensitive files, network operations, or shell command execution.
- Sanitization: No input filtering or sanitization is performed on the provided review text before it is interpolated into the agent's prompt.
Audit Metadata