triggers
Pass
Audited by Gen Agent Trust Hub on Aug 29, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill defines triggers (Email, HTTP, Kafka, etc.) that ingest untrusted data from external sources, which serve as ingestion points for headers, bodies, and attachments. The configuration lacks mandatory boundary markers to isolate this untrusted data from instructions. The associated scripts or flows (runnables) have broad capabilities, creating a potential surface for indirect prompt injection if the ingested data contains malicious instructions. No sanitization or validation logic is defined for these inputs in the trigger schemas.
- [COMMAND_EXECUTION]: The skill references the 'wmill sync push' and 'wmill sync pull' CLI commands for synchronizing trigger configurations. It includes a specific instruction for the agent to only suggest or run deployment commands when the user explicitly requests to deploy, acknowledging that these operations can be destructive to remote state.
Audit Metadata