cross-sell-upsell-engine

Pass

Audited by Gen Agent Trust Hub on Mar 16, 2026

Risk Level: SAFE
Full Analysis
  • [DATA_EXPOSURE_AND_EXFILTRATION]: The skill includes template code for fetching recommendation data from Rebuy Engine. This is a well-known third-party e-commerce personalization service.
  • Evidence: Fetch call to https://rebuyengine.com/api/v1/products/recommended using a placeholder API key.
  • [INDIRECT_PROMPT_INJECTION]: The skill defines a surface for processing untrusted order data to generate product recommendations.
  • Ingestion points: Order history is ingested from the database via db.orders.findAll in SKILL.md.
  • Boundary markers: Not explicitly implemented in the provided architectural templates.
  • Capability inventory: Includes network capabilities to fetch data from external APIs and file-write capabilities for affinity records.
  • Sanitization: Data is processed for statistical affinity without explicit sanitization, which is standard for recommendation logic.
  • [SAFE]: No obfuscation, unauthorized command execution, or persistence mechanisms were detected across the skill files.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 16, 2026, 12:47 PM