native-trigger
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill provides a framework for implementing webhook-based triggers that ingest data from external services (e.g., Google Drive, Nextcloud) to execute scripts or flows. This creates an inherent attack surface for indirect prompt injection where malicious payloads from external services could influence the behavior of triggered scripts.\n
- Ingestion points: Webhook request bodies and headers processed in the
prepare_webhookmethod (referenced inSKILL.md).\n - Boundary markers: The implementation guide does not explicitly include or mandate the use of delimiters or instructions to ignore embedded commands in the ingested data.\n
- Capability inventory: The system is designed to trigger Windmill scripts and flows, which may have broad permissions to interact with the filesystem, network, or databases depending on their configuration.\n
- Sanitization: The provided code templates demonstrate direct mapping of JSON payload fields to script arguments without explicit sanitization or validation logic, relying on the downstream script's own handling.
Audit Metadata