What you'll learn
Design and implement a production-grade CLAUDEmd that encodes architecture rules, guardrails, testing policies, and governance standards.
Use Plan Mode to transform ambiguous feature ideas into structured artifacts (REQUIREMENTSmd, ROADMAPmd, STATEmd) with phased execution.
Build and deploy reusable Claude Skills and custom slash commands to automate real-world developer workflows.
Implement mandatory test gating, CI enforcement, and human-in-the-loop workflows for safe AI-assisted development.
Configure Claude Code for large repositories and monorepos using layered governance and folder-level overrides.
Detect and mitigate AI failure modes such as hallucinations, unsafe migrations, and silent logic drift.
Design and document an organizational safe usage policy for enterprise-scale AI adoption.
Deliver a complete capstone repository demonstrating AI-native development practices from planning to production governance.

Requirements
Strong proficiency in at least one programming language (e.g., Python, javascript/TypeScript, Java, Go, etc.).
Experience working with Git and modern pull request workflows.
Familiarity with backend or full-stack architecture (APIs, services, databases, CI/CD basics).
Understanding of automated testing concepts (unit tests, integration tests, CI pipelines).
Comfort working in a command-line environment and navigating project repositories.
Basic knowledge of software design principles (modularity, separation of concerns, versioning).
Access to a development machine capable of running Claude Code (CLI or Desktop) and a sample repository to practice on.
Recommended (not mandatory): Experience working in a team-based development environment or monorepo setup.