amz-review-analyzer

Pass

Audited by Gen Agent Trust Hub on Jun 25, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill is purely instructional and does not define any tools, network operations, or file system access. It functions as a set of logical steps for the AI to follow when processing text provided by the user.
  • [PROMPT_INJECTION]: The skill involves processing untrusted data from user-pasted reviews. 1. Ingestion points: User-provided product reviews mentioned in SKILL.md. 2. Boundary markers: Absent. 3. Capability inventory: No tools or subprocess capabilities are defined or used. 4. Sanitization: Absent. The lack of any functional tools or system access ensures that even if a review contained malicious instructions, the agent would have no means to execute harmful actions.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 25, 2026, 01:08 PM
Security Audit — agent-trust-hub — amz-review-analyzer