sent-webhook-engineer
Pass
Audited by Gen Agent Trust Hub on Oct 7, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill provides instructions for processing untrusted webhook payloads from external sources. This represents a vulnerability surface if the ingested data is later used in sensitive contexts.
- Ingestion points: Framework recipes in
references/receiver-recipes.mdand the catalog inreferences/event-catalog.mddescribe handlers that ingest external JSON payloads via HTTP POST. - Boundary markers: The implementation relies on HMAC signature verification via
x-webhook-signaturefor authenticity, but does not define delimiters for payload data if used in downstream prompts. - Capability inventory: The skill includes a standalone Python oracle (
scripts/verify_signature.py) and various framework code snippets for signature verification and background processing. - Sanitization: The
SKILL.mdfile provides a mandatory security boundary instruction, stating that payload values must never be interpolated into shell commands, SQL strings, or prompts without escaping.
Audit Metadata