product-feedback-synthesizer

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

• Third-party content exposure detected (high risk: 0.90). The SKILL.md explicitly requires automated collection and ingestion of user-generated content from public channels (e.g., "social media monitoring", "community forums", "Discord, Reddit", and "review sites" in the Collection Strategy and Processing Pipeline), meaning the agent will read and act on untrusted third‑party content that could carry indirect prompt injections.