incident-sentiment
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze untrusted data from ServiceNow's
sys_journal_field(work notes and comments) andsys_email(customer communications). - Ingestion points: Untrusted content is retrieved in Step 2 (journal entries) and Step 3 (emails) using the
SN-Query-Tabletool and REST API calls. - Capability inventory: The agent possesses the capability to write back to the system, modify records, and change incident urgency via
SN-Update-RecordandSN-Add-Work-Notestools. - Boundary markers: The skill instructions do not specify any boundary markers or delimiters to separate instructions from the data being analyzed, nor does it provide directives for the agent to ignore instructions embedded within the text data.
- Sanitization: While Step 4 and the troubleshooting section mention stripping HTML or analyzing keywords, there is no explicit validation or sanitization process to prevent instructions hidden in the text from being executed by the agent.
Audit Metadata