agent-workspace
Pass
Audited by Gen Agent Trust Hub on Sep 6, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill provides standard ServiceNow GlideRecord and UI Action templates for building Agent Workspaces. All code snippets follow documented ServiceNow development practices and Serac-Labs patterns.
- [SAFE]: Tools such as snow_query_table and snow_execute_script are utilized according to their primary purpose of querying ServiceNow instances and testing server-side scripts during development.
- [INDIRECT_PROMPT_INJECTION]: The skill processes record data from ServiceNow tables (e.g., incident details, workspace configurations). This represents a standard data ingestion surface for development tools and is managed through standard platform filters and GlideRecord queries. Ingestion points: Data retrieved from sys_aw_* tables and incident records via snow_query_table and custom widget scripts. Boundary markers: Standard ServiceNow data structures are used; instructions are not interpolated into prompts. Capability inventory: The skill uses tools for querying tables, managing artifacts, and executing server-side scripts on a ServiceNow instance. Sanitization: Example scripts include conditional checks (e.g., if (tableName === 'incident')) to ensure scripts operate on expected record types.
Audit Metadata