customer-feedback-triage

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

• Third-party content exposure detected (high risk: 0.75). The required runtime workflow ingests outsider-authored customer feedback verbatim (raw_text) from an operator-supplied JSON queue (python scripts/feedback_triage.py --input queue.json), and the tool then renders that text into LLM-readable output (e.g., Markdown/Confluence/Notion response drafts), so any free-form customer text can reach the agent’s LLM context indirectly via the generated report.