A serverless CRM on AWS, a multi-agent assistant built on Strands Agents, and two MCP patterns on Amazon Bedrock AgentCore Gateway. Together they form a measurement instrument. The same agentic tasks run against Anthropic and OpenAI models on Amazon Bedrock. The harness records correctness, write safety, tool use and token counts for each.
evaluations mcp crm cognito bedrock eval evaluation-framework genai evals genai-poc genai-security agentic-ai mcp-servers mcp-server evaluation-harness bedrock-agentcore eval-harness agentcore-gateway genai-models agentcore-runtime
-
Updated
Nov 16, 2026 - Python