{"success":true,"data":{"id":"rag-corpus-poisoning-via-mail","label":"Email from unknown sender presenting itself as official company policy / documentation update designed to be ingested into a RAG knowledge base. Claims authority (\"supersedes all previous guidance\", \"effective immediately\") and either explicitly instructs ingestion (\"add to your RAG corpus / knowledge base\") or combines policy-authority language with knowledge-base vocabulary. Label-only: engine cannot block RAG ingestion but must surface the email and refuse silent delete. Source: Red-Team R8 multi-agent council C5 (agentic-AI specialist).","description":"Email from an unknown sender (no List-Unsubscribe) presenting itself as official company policy or documentation update designed to be ingested into a RAG knowledge base. Claims authority (\"supersedes all previous guidance\", \"effective immediately\", \"overrides existing documentation\") and either explicitly instructs ingestion (\"please add to your RAG corpus / knowledge base\") or combines policy-authority language with knowledge-base vocabulary. The attack goal is to pollute the RAG knowledge base so future AI-assistant queries return attacker-controlled guidance. Label-only signal (invoice +5): engine cannot block RAG ingestion but must surface the email and refuse silent delete. Source: Red-Team R8 multi-agent council C5 (agentic-AI specialist).","tier":"warning","category":"other","isThin":false}}