clean-code

Apply explicit rules for naming, function size, and structure such as SRP, DRY, KISS, and YAGNI. This Skill encodes those rules as mandatory instructions so the AI writes concise, self-documenting code without obvious comments or over-engineering.

Core principles include Single Responsibility, Don't Repeat Yourself, Keep It Simple, and You Aren't Gonna Need It. Functions should stay under 20 lines, take at most 3 arguments, and use guard clauses instead of deep nesting.

Before editing any file, check what imports it and what it imports, then update all dependent files in the same task. This Skill mandates that dependency check as a required pre-edit step.

Models default to tutorial-style output explaining each line. The fix is instructing the AI to let code self-document through clear naming and to delete comments that restate obvious logic.

Run the validation script matching the agent's specialty immediately after completing work, then summarize errors and warnings to the user before fixing. Auto-fixing without user confirmation is not allowed.