paystack-testing
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [SAFE]: The skill follows security best practices by providing placeholders for API keys and explicitly warning against committing real secrets to version control.
- [INDIRECT_PROMPT_INJECTION]: The skill defines a surface for processing external data through webhook handlers. This is addressed with the following considerations:
- Ingestion points: External webhook payloads processed in
webhookHandler(SKILL.md) and API responses inpaystackRequest(SKILL.md). - Boundary markers: The implementation lacks explicit natural language boundary markers but relies on structured JSON processing.
- Capability inventory: Uses
fetchfor network requests and Node.jscryptofor signature verification. - Sanitization: Includes a robust signature validation check using HMAC SHA512 to ensure data integrity and authenticity from Paystack.
- [EXTERNAL_DOWNLOADS]: Mentions well-known developer services
ngrokandlocaltunnelfor local development testing. These are standard industry tools for exposing local environments and do not represent a security risk in this context.
Audit Metadata