Research artifacts exploring Search-based vs. Agentic RAG strategies for AI Code Review. A deep dive into solving the "Context Instability" problem in LLM-based software engineering, analyzing trade-offs in recall, latency, and cost.