buy-agent-commerce-guard

Pass

Audited by Gen Agent Trust Hub on Jul 23, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADS
Full Analysis
  • [SAFE]: The skill incorporates explicit safety instructions that strictly forbid the agent from requesting sensitive information such as private keys, seed phrases, or wallet recovery codes.
  • [SAFE]: Implements a mandatory validation step where the agent must verify that the transaction network, token contract, receiver address, and payment amount match hardcoded constants before proceeding.
  • [EXTERNAL_DOWNLOADS]: Fetches product metadata, licensing information, and payment configurations from the vendor's official website hosted on Vercel (agent-commerce-guard.vercel.app).
  • [SAFE]: The skill enforces an approval-gated workflow, ensuring that no payment or wallet action is initiated without explicit user confirmation.
  • [SAFE]: The ingestion of external JSON metadata and tool catalogs from the vendor's domain is managed through specific validation logic that mitigates the risk of indirect prompt injection.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 23, 2026, 07:29 PM
Security Audit — agent-trust-hub — buy-agent-commerce-guard