What problem does it solve?
Enforce a single, high-signal engineering standard across codebases so reviewers and authors produce readable, robust, maintainable, and type-safe code. The skill collapses many recurring review complaints (ambiguous names, brittle control flow, hidden side-effects, silent failures, and leaky DB/entity shapes) into concrete, automatable checks and actionable guidance.
Core Features & Use Cases
- Comments & Why — teach reviewers to write explanatory comments that document consequences, invariants, and limits rather than paraphrase code.
- Naming & Semantics — prefer intentful names, symmetry, explicit units, and ban vague verbs and single-letter identifiers.
- Control Flow & Functions — guard clauses, no boolean flags, Promise.all for independent async ops, one level of abstraction per function, and factory DI.
- Data, Types & Errors — immutable data, strict typing, bound every input at the barricade, Result/Option patterns over exceptions, and structured API errors with DTO mapping.
- File Structure & Architecture — workflow-first files, locality of behavior, vertical feature slices, DI factories, and crosscutting via wrappers (withLogging/withTracing).
- Project Hygiene — externalize config, enforce lint/tests/CI early, remove dead code, and make invalid states unrepresentable.
- Use Cases: run on PR reviews, refactors, architecture proposals, and as a checklist for writing new features or onboarding codebases.
Quick Start
Ask the assistant to review a file or pull request for adherence to these coding standards and return a prioritized list of concrete fixes for naming, control flow, functions, types, error handling, comments, file structure, architecture, and hygiene.
