Test and Refine Agent Collaboration

Create a test suite that simulates a stream of diverse prompts, including those with mixed sentiment and specific brand mentions. Verify that the `PromptMonitor` correctly ingests them, the `BrandAnalyzer` accurately applies DeepSeek R1 for sentiment, and the `ReportGenerator` produces coherent summaries. Pay close attention to the handover of information between agents. Use Blaxel for observing agent interactions and potential bottlenecks. Provide examples of expected intermediate agent outputs at each stage.