name: code-quality description: Establish enterprise-grade code quality baselines for the .NET solution: analyzers, nullability, formatting, complexity limits, and review-ready diffs. Use when changing production code, refactoring, or introducing new project patterns.

Code Quality

Overview

Turn “good engineering habits” into repeatable quality gates.

This skill is intentionally cross-cutting: it applies to UI, application, domain, and infrastructure code.

When to use

• Any non-trivial code change (new feature, bug fix, refactor)
• Adding a new project, package, or architectural pattern
• Touching error-prone areas (async, threading, I/O, serialization)

Definition of done (DoD)

• Builds cleanly locally (dotnet build) with no new warnings introduced
• Tests pass for the nearest scope (dotnet test for the affected test project(s))
• Nullability decisions are explicit (no “spray-and-pray” ! suppression)
• Public APIs are intentional (avoid accidental public surface)
• Changes are small and reviewable (prefer focused commits/diffs)

Practical checklist

Readability and maintainability

• Prefer descriptive names over abbreviations
• Avoid deep nesting; extract methods when cognitive load increases
• Keep side effects localized (especially in constructors)

Error handling

• Throw exceptions for programming errors; return results for expected failures
• Do not swallow exceptions; either handle and continue safely, or log/report and fail fast

Async and threading

• Avoid blocking (.Result, .Wait()) on async paths
• Prefer CancellationToken for long-running/interactive operations
• Keep UI thread work minimal; offload I/O and CPU-bound work

Analyzers/formatting (repo-level standards)

• If .editorconfig is missing and you’re doing broad work, consider adding one alongside the change
• Prefer enabling analyzers over suppressing them
• If a suppression is necessary, scope it narrowly and document the reason

Common anti-patterns

• Catch-all catch (Exception) without a clear recovery strategy
• Logging secrets/PII or large payloads
• “Fix” via disabling warnings globally instead of addressing root cause

Notes

This skill describes quality gates; the exact enforcement mechanism (CI pipeline, analyzers, formatting tooling) should evolve over time.