project-guidelines-example

Copy this template and replace its sections with your own architecture overview, file structure, code patterns, testing requirements, and deployment workflow. Keep the SKILL.md frontmatter with a unique name and description so the skill is discoverable.

Include the tech stack, service architecture diagram, directory structure, canonical code patterns for APIs and components, testing commands and structures, deployment steps, and critical rules such as coverage minimums and immutability constraints.

Define a Pydantic model for the result, pass its JSON schema as a tool definition, and set tool_choice to force that tool. The response contains a tool_use block whose input validates directly against your model.

Yes, the structure is stack-agnostic. Swap the example patterns for your own framework conventions while keeping the same sections: architecture, file structure, code patterns, testing, deployment, and critical rules.

Explicit critical rules constrain AI-generated code to match team standards, preventing common issues like mutated state, oversized files, missing error handling, and console.log statements in production code.