hrsd-sentiment-analysis

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

• Third-party content exposure detected (medium risk: 0.30). Runtime workflow fetches and LLM-analyzes free-text fields from outsider-originated sources stored in first-party tables—HR case work notes/comments (sys_journal_field.value) and interaction wrap-up/comments (interaction.wrap_up_comment, etc.) for sentiment scoring, so an outsider can inject text into those channels and it will be ingested during the case/communication extraction steps.