aws-solutions-library-samples

Scaffold a new A2A capability agent with Python application, Dockerfile, requirements.txt, and CDK stack following the project's established patterns

Scaffold a new local tool for the voice agent pipeline with capability-based registration, executor function, ToolDefinition, and tests

Run an ECS scaling validation test with parallel monitoring, baseline enforcement, and automated result documentation

Sets up Daily.co phone number and webhook for PSTN dial-in. Guides through API key verification, phone number purchase, pinless dial-in configuration, and secrets sync. Use after deploying infrastructure, when setting up a phone number, or when configuring dial-in.

Scaffolds a new A2A capability agent with Python application, Dockerfile, requirements.txt, and CDK stack. Use when adding a new remote tool or service that the voice agent discovers via CloudMap.

Scaffolds a new local tool for the voice agent pipeline with capability-based registration, executor function, and tests. Use when adding a tool that runs inside the voice agent container and may need transport or SIP session access.

Deploys optional A2A capability agents (Knowledge Base, CRM) that extend the voice agent with new skills. Explains the A2A architecture, enables the capability registry, deploys agent stacks, and verifies discovery. Use after the core deployment and Daily setup are complete.

Deploys the voice agent to AWS using Deepgram and Cartesia cloud APIs. Guides through prerequisites, environment setup, CDK deployment, and secrets configuration. Use when deploying for the first time, setting up a dev environment, or when SageMaker is not needed.

Organization code review checklist

Implement GRPO (Group Relative Policy Optimization) fine-tuning for vision-language models on small datasets. Use when SFT underperforms or training data is limited (<1000 examples).

Use Bedrock tool_use to guarantee structured JSON outputs from Claude models. Eliminates JSON parsing failures by forcing responses through typed tool schemas.

Implement SageMaker async inference with S3-based I/O and polling. Use for long-running inference (>60s), large payloads, or batch processing workloads.

Create distinctive, production-grade frontend interfaces with high design quality. Use this skill when the user asks to build web components, blogs, pages, or applications. Generates creative, polished code that avoids generic AI aesthetics.