scheduled-jobs
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEPERSISTENCEDYNAMIC_EXECUTIONDATA_EXFILTRATIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [PERSISTENCE]: The skill provides detailed instructions and templates for creating
sysauto_script(Scheduled Script Executions) andsys_trigger(System Triggers) records. In the ServiceNow ecosystem, these records are the primary and intended mechanism for establishing persistent, recurring background execution of logic. - [DYNAMIC_EXECUTION]: The skill utilizes tools to execute logic via
SN-Execute-Background-Scriptand demonstrates advanced patterns wheresys_triggerrecords are programmatically created with JavaScript code strings to facilitate self-scheduling and job dependency chains. - [DATA_EXFILTRATION]: The skill includes code examples for sending outbound data via the
GlideEmailOutboundAPI andsn_ws.RESTMessageV2for integration purposes. These patterns utilize placeholder domains and email addresses (e.g.,api.example.com,company.com) to demonstrate functional reporting and synchronization capabilities. - [INDIRECT_PROMPT_INJECTION]: The skill demonstrates patterns for ingesting data from various ServiceNow tables which is then interpolated into logs, email notifications, and outbound JSON payloads.
- Ingestion points: Untrusted data enters the agent context through
GlideRecordandGlideAggregatequeries performed on tables such asincident,sys_user, and custom integration queues. - Boundary markers: The provided script templates do not employ specific delimiters or boundary markers when processing record data.
- Capability inventory: The skill leverages
SN-Create-Record,SN-Update-Record, andSN-Execute-Background-Scriptto perform record operations and execute administrative logic based on ingested data. - Sanitization: The examples demonstrate standard use of
JSON.stringify()for structured logging but do not include explicit input validation or sanitization of record field values before they are used in outbound communications.
Audit Metadata