ai-coding-discipline

To stop AI generated code from swallowing errors, enforce proper error handling rules that prohibit catch-all try/catch blocks in business logic, allowing errors to propagate naturally instead of being masked.

The fail-fast principle for AI coding prevents silent null fallbacks by mandating that code throws errors immediately when encountering invalid data, exposing data issues early rather than hiding them behind default values.

Writing meaningful tests for bug fixes using red-green TDD requires mandating tests that explicitly fail when logic breaks, ensuring new code accurately satisfies exact business outcomes like discount calculations.

Yes, this coding discipline skill applies directly to refactoring and code review tasks, enforcing strict rules across writing, editing, debugging, and testing to prevent common AI coding anti-patterns.

Common AI coding anti-patterns when implementing features include hiding bugs with silent fallbacks, swallowing errors with catch-all blocks, and creating weak tests that fail to verify exact business outcomes.

You should load coding discipline rules before writing or editing any code in AI-assisted development, ensuring fail-fast principles, proper error propagation, and strong testing practices are enforced from the start.