case-summarization-approvals

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (medium risk: 0.30). At runtime the workflow fetches and LLM-ingests outsider-authored free text stored in ServiceNow records—e.g., Step 3 extracts justification from catalog variable question_text/value and Step 3 also pulls journal/comments from sys_journal_field for the source case/request, then Step 5 pulls approver comments.