recommendations

Pass

Audited by Gen Agent Trust Hub on Sep 16, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill extracts user preferences, psychological profiles, and history from untrusted conversation data to build context for remote API requests.
  • Ingestion points: User conversation extracted into JSON context fields in SKILL.md (Step 1) and references/context-patterns.md (Preference Extraction).
  • Boundary markers: Absent; user-derived content is directly interpolated into JSON structures sent to the API.
  • Capability inventory: Network requests (POST /v1/recommend, POST /v1/explain) to api.tasteray.com as documented in references/api-reference.md.
  • Sanitization: The instructions do not specify any sanitization, filtering, or validation for user-provided preference or profile strings.
  • [EXTERNAL_DOWNLOADS]: The skill performs network operations to interact with the TasteRay service.
  • Details: All documented API calls target api.tasteray.com. This domain corresponds to the skill author's own infrastructure (tasteray) and is essential for the skill's core functionality of providing recommendations.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 16, 2026, 09:00 AM
Security Audit — agent-trust-hub — recommendations