predict-assignment
Pass
Audited by Gen Agent Trust Hub on May 20, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill's stated purpose of predicting incident assignments is consistent with its use of ServiceNow MCP tools (SN-NL-Search, SN-Query-Table, etc.) and REST API endpoints.
- [PROMPT_INJECTION]: The skill exhibits a surface for indirect prompt injection as it processes untrusted incident descriptions. However, it includes mitigations such as confidence scoring thresholds and manual triage requirements for ambiguous inputs, which are considered safe practices for this use case.
- [DATA_EXFILTRATION]: Data access is limited to standard ServiceNow tables required for ITSM workflows. There is no evidence of sensitive data being transmitted to external or unauthorized domains.
- [COMMAND_EXECUTION]: While Bash is listed as a native tool, its usage in the documentation is limited to providing examples of standard REST API calls (GET/PATCH) for integration purposes.
Audit Metadata