modify-agent-prompts

To modify agent prompts and tool configuration, you adjust the review-item-processor prompts, model IDs, and tool usage instructions. This process enforces guidelines for prompt structure, JSON schema formatting, and environment variable references to tailor AI agent behavior.

The best way to update model IDs and tool instructions is to apply configuration governance during deployment. This ensures safe defaults, documents changes, and guides structured rollouts for production systems while maintaining consistent agent capabilities across environments.

Yes, you can configure agent prompts and tool usage instructions using CDK. The Skill supports agent-configuration tasks during development and deployment, allowing you to update tool configurations and enforce environment variable references within your infrastructure code.

When tuning agent prompts, you must follow enforced guidelines for JSON schema formatting and environment variable references. Proper structure ensures the review-item-processor correctly interprets tool configurations and executes tailored AI agent behavior without formatting errors.

When adding or removing tools for the review-item-processor, you are constrained by enforced guidelines for prompt structure and agent architecture notes. You must document changes and ensure safe defaults to maintain stable production deployments during tool modifications.