software-payments
Pass
Audited by Gen Agent Trust Hub on Mar 12, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [SAFE]: The skill follows security best practices by recommending environment variable storage for sensitive API keys and secrets rather than hardcoding them. It also correctly implements lazy-initialization to prevent build-time failures.
- [PROMPT_INJECTION]: Analyzed the attack surface for indirect prompt injection via external payment webhooks.
- Ingestion points: External data from payment providers (Stripe, Adyen, etc.) is ingested through webhook endpoints in
SKILL.md. - Boundary markers: The skill explicitly requires using cryptographic signature verification (e.g.,
stripe.webhooks.constructEvent) to authenticate incoming data. - Capability inventory: The skill triggers database updates for subscription states, handles entitlement gating, and processes user-level metadata changes.
- Sanitization: Mitigations include UUID validation for user IDs within metadata and strict signature verification for all incoming payloads.
Audit Metadata