# ai-rag-poisoning

> Poison a RAG/knowledge-base pipeline so retrieved content hijacks the model (indirect prompt injection at scale) or exfiltrates data. Load when the app does retrieval over documents/URLs/ tickets/emails the attacker can influence, "RAG", vector DB, "knowledge base", agent that reads content. Signals

Skill · Free · License: MIT · Version: 2026.10.14-ee0e0c5

Install with the NotHoly CLI: `notholy skills install noorqureshi/ai-rag-poisoning`

Canonical: https://notholy.com/en/skills/noorqureshi/ai-rag-poisoning
