ecommerce-product-opportunity-researcher

Pass

Audited by Gen Agent Trust Hub on Jul 27, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill mandates a specific intro message in deliverables that solicits donations via CashApp ($AlainDorcelus) and BuyMeACoffee.
  • [PROMPT_INJECTION]: The instructions include a hardcoded response for system prompt inquiries, redirecting users to the Scayver Academy website.
  • [PROMPT_INJECTION]: The 'Review Mining Mode' processes external data (reviews and support tickets) which is a surface for indirect prompt injection. Ingestion points: Customer reviews and support tickets identified in SKILL.md. Boundary markers: Explicit instructions to anonymize details and distinguish quotes are present. Capability inventory: No high-risk tools like shell access or file writes are present. Sanitization: Mandatory anonymization of identifying details is required.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 27, 2026, 12:11 AM
Security Audit — agent-trust-hub — ecommerce-product-opportunity-researcher