field-service
Pass
Audited by Gen Agent Trust Hub on Aug 31, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill defines templates for ServiceNow scripts that interpolate data from external variables (e.g.,
workOrderSysId,tasks[i].description,statusUpdate.notes) intoGlideRecordoperations. If the agent processes untrusted data as input for these fields, there is a minor risk that specially crafted content could influence the logic of the generated scripts or the data written to the ServiceNow instance. - Ingestion points: Inputs to
createWorkOrderTasks,updateWorkOrderFromMobile, andrecordTimeEntryinSKILL.md. - Capability inventory: Uses
snow_execute_scriptto run server-side JavaScript on ServiceNow. - Sanitization: The provided snippets do not show explicit input validation or escaping before setting values via
setValueorwork_notes. - [COMMAND_EXECUTION]: The skill relies on the
snow_execute_scripttool to perform administrative and operational tasks within a ServiceNow environment. This is a functional requirement for managing FSM workflows but grants the agent the ability to execute code within the target ServiceNow instance's scope.
Audit Metadata