frontpage-buy-ad

Pass

Audited by Gen Agent Trust Hub on Jun 24, 2026

Risk Level: SAFEDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill retrieves ad data from the frontpage.sh API, which includes content submitted by other users, creating a risk of indirect prompt injection.\n
  • Ingestion points: Data enters the agent context via the GET /api/ads endpoint, which returns fields like ctaLabel, perk, and promoCode populated by previous ad buyers.\n
  • Boundary markers: The instructions do not define delimiters or specific warnings to ignore embedded instructions within the fetched content.\n
  • Capability inventory: The skill can perform network requests and execute financial transactions (USDC payments via MPP).\n
  • Sanitization: The process relies on server-side moderation rather than agent-side validation or sanitization of ingested data.\n- [DATA_EXFILTRATION]: The skill collects and transmits the user's email address (ownerEmail) to the frontpage.sh API for receipts and notifications. While part of the stated functionality, this involves the transmission of personally identifiable information (PII) to an external service.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 24, 2026, 09:24 PM
Security Audit — agent-trust-hub — frontpage-buy-ad