The Agent Skills Directory

[DYNAMIC_EXECUTION]: The skill uses the SN-Execute-Background-Script tool to run ServiceNow server-side JavaScript. It constructs these scripts by directly interpolating variables (such as [ritm_sys_id], [cat_item_sys_id], and [department_sys_id]) into the script body. This practice of building executable code via string concatenation of variable inputs introduces a risk of code injection if the platform does not properly sanitize these inputs.\n- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from ServiceNow records, creating a surface for indirect prompt injection.\n
Ingestion points: Data is retrieved from fields like business justification, comments, and short_description in the sysapproval_approver, sc_req_item, and sc_item_option_mtom tables.\n
Boundary markers: The skill lacks explicit markers or instructions to prevent the agent from following commands embedded within the fetched data.\n
Capability inventory: The skill has access to sensitive capabilities including arbitrary script execution (SN-Execute-Background-Script), record updates (SN-Add-Work-Notes), and local shell commands (Bash).\n
Sanitization: There is no evident sanitization of the record data before it is incorporated into prompts for summarization and decision support.

approval-summarization