interview-relevance

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

• Third-party content exposure detected (medium risk: 0.30). At runtime, the workflow queries and then parses free-text fields like u_description (job description) and u_notes/u_strengths/u_concerns (interview feedback) from ServiceNow tables via SN-Query-Table/REST, so an outsider who can post malicious text into those records can influence the LLM before any specific selection step.