Rag Pipeline Debugger

This prompt acts as a senior AI engineer who diagnoses your retrieval-augmented generation pipeline. Paste in your user query, retrieved chunks, and the model's final response and get back a structured fault report covering: context utilisation failures, hallucination triggers, retrieval mismatches, prompt injection risks, and a prioritised fix plan with exact prompt rewrites. Works with any RAG setup: custom pipelines, vector DBs. Saves at least 4 hours of manual debugging per session.