academic-pipeline

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

• Third-party content exposure detected (low risk: 0.10). The workflow reads only the user’s own conversation input for intent detection and checkpoints, and the only outsider-authored free text it “audits” (Stage 4→5 claim alignment) is the Stage 4 draft it previously generated/received within the pipeline, not arbitrary external submissions via queue/feed.