Webhook Receiver Hardener
Pass
Audited by Gen Agent Trust Hub on Sep 5, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill follows security best practices for webhook handling. It provides defensive measures against timing attacks, replay attacks, and request tampering.
- [INDIRECT_PROMPT_INJECTION]: The skill documents the handling of untrusted external data (webhooks). It correctly identifies this as a hostile-input surface and provides clear boundaries for mitigation:
- Ingestion points: The
SKILL.mdidentifies inbound webhook routes (e.g.,/webhooks/stripe) as the entry point for untrusted data. - Boundary markers: The skill mandates HMAC signature verification over raw request bytes using constant-time equality checks before any data processing occurs.
- Capability inventory: The skill outlines a limited pipeline of
verify→persist→enqueue→ack, specifically instructing the agent not to run business logic or blocking operations within the request context. - Sanitization: The instructions require that
JSON.parseand business logic only occur after successful cryptographic verification.
Audit Metadata