problem-management

Pass

Audited by Gen Agent Trust Hub on Feb 23, 2026

Risk Level: SAFE
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses the tool snow_execute_script_with_output to perform server-side automation, such as proactive problem identification based on incident patterns. This behavior is consistent with the primary purpose of a ServiceNow development skill.
  • [PROMPT_INJECTION]: The skill includes search functionality that ingests incident descriptions to query the Known Error Database. While this creates a surface for indirect data processing, it is used for legitimate search operations within the ServiceNow environment and does not attempt to bypass agent safety filters.
  • [DATA_EXFILTRATION]: The skill interacts with ServiceNow tables such as problem, incident, and known_error. These operations are conducted using standard GlideRecord APIs for the purpose of record management and do not involve unauthorized data transfer to external domains.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 23, 2026, 08:31 PM