change-rollback

Pass

Audited by Gen Agent Trust Hub on Sep 6, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill provides instructions for the agent to ingest and process payloads from the sys_update_version table in ServiceNow. This creates a surface where a malicious payload stored in a ServiceNow record could attempt to influence the agent's behavior during the restoration process.\n
  • Ingestion points: Payloads from the sys_update_version table, retrieved via snow_query_table.\n
  • Boundary markers: The instructions guide the agent to parse XML tags and CDATA sections, providing structural boundaries, though the skill does not mandate semantic validation of the scripts being restored.\n
  • Capability inventory: The skill uses snow_artifact_manage for record updates and snow_execute_script for server-side JavaScript execution.\n
  • Sanitization: The skill explicitly advises on decoding XML entities and unwrapping CDATA to maintain data integrity during the restore operation.\n- [DYNAMIC_EXECUTION]: The skill references the snow_execute_script tool, which is used to execute arbitrary ES5 JavaScript on a ServiceNow instance. While suggested only as a fallback for specific API limitations, this functionality involves runtime code execution that requires careful handling.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 6, 2026, 07:16 PM
Security Audit — agent-trust-hub — change-rollback