The Agent Skills Directory

[DYNAMIC_EXECUTION]: The skill relies on the SN-Execute-Background-Script tool to run ServiceNow-specific JavaScript (GlideRecord and GlideAggregate) directly on the ServiceNow instance. These scripts are provided as templates within the skill's instructions. While standard for complex ServiceNow automation, this practice involves dynamic execution of embedded code at runtime. A risk exists if the AI agent interpolates user-controlled parameters, such as [service_sys_id] or [ci_sys_id], into these scripts without strict validation, potentially enabling script injection on the target ServiceNow platform.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and act upon data from the ServiceNow CMDB, creating a potential attack surface for indirect prompt injection.
Ingestion points: The agent retrieves data from multiple ServiceNow tables, including cmdb_ci_service, svc_ci_assoc, sa_pattern, and em_alert using the SN-Query-Table and SN-Get-Record tools.
Boundary markers: The instructions lack explicit boundary markers or instructions to treat data retrieved from the CMDB as untrusted content.
Capability inventory: The skill grants the agent access to high-impact tools, specifically SN-Execute-Background-Script for remote execution and SN-Update-Record for modification of CMDB records.
Sanitization: There are no requirements or logic provided for sanitizing or filtering data fetched from the external ServiceNow tables before it is processed by the agent.

service-mapping