Here is a brief, actionable plan for testing long-context prompt chunking, structured in four phases.

**Goal:**