webhook-processing
Pass
Audited by Gen Agent Trust Hub on Sep 25, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill describes an architecture for ingesting untrusted data from external email provider webhooks (e.g., SendGrid, Postmark, AWS SES), which presents an attack surface for indirect prompt injection.
- Ingestion points: Webhook request bodies and headers processed at HTTP endpoints (e.g.,
/webhooks/resend) as detailed inSKILL.md. - Boundary markers: The skill focuses on system-level verification and normalization but does not provide specific instructions for sanitizing content before potential interpolation into LLM prompts.
- Capability inventory: The skill examples include database writes (
db.query), job queueing (queue.add), and recording audit events. - Sanitization: Recommends mandatory cryptographic signature verification (HMAC-SHA256, ECDSA) to authenticate the data source.
- [SAFE]: The skill follows security best practices by recommending
timingSafeEqualfor signature comparisons to mitigate timing attacks. - [SAFE]: The skill identifies the risk of SSRF when handling AWS SNS subscription confirmation requests and provides a hostname validation check to ensure the URL originates from legitimate AWS domains.
Audit Metadata