Agent skill

Domain-Focused Naming

Name code by what it does in the domain, not how it's implemented or its history

Stars 163
Forks 31

Install this agent skill to your Project

npx add-skill https://github.com/majiayu000/claude-skill-registry/tree/main/skills/development/domain-focused-naming

SKILL.md

Domain-Focused Naming

Overview

Names documenting implementation or history create confusion. "NewUserAPI" doesn't tell what it does. "ZodValidator" exposes internals.

Core principle: Names tell what code does in the domain, not how it's built or what it replaced.

Violating the letter of this rule is violating the spirit of naming.

When to Use

Use for:

  • Variables, functions, classes, modules
  • Refactoring existing code
  • Code review feedback
  • API design

Use ESPECIALLY when:

  • Refactoring (tempted to add "New" or "Improved")
  • Replacing implementations (tempted to add "Zod" or "MCP")
  • Using design patterns (tempted to add "Factory" or "Manager")
  • Documenting changes (tempted to add "Unified" or "Enhanced")

The Rules

NEVER Use Implementation Details

Names expose WHAT, not HOW.

NEVER Use Temporal Context

Code exists in present. Don't reference past or transitions.

NEVER Use Pattern Names (Unless They Add Clarity)

Patterns are implementation details. Most don't help understanding.

// OK when pattern IS the purpose class EventEmitter { } // Observer pattern IS what it does class CommandQueue { } // Queue pattern IS what it does

</Good>

### Names Tell Domain Stories

Good names form sentences about business logic.

<Good>
```typescript
// Reads like domain language
user.authenticate()
order.calculateTotal()
payment.process()

// Not
user.executeAuthenticationStrategy()
order.runTotalCalculationAlgorithm()
payment.invokeProcessingWorkflow()

Quick Reference

Bad Pattern Why Bad Good Alternative
ZodValidator Exposes implementation Validator
MCPToolWrapper Exposes protocol RemoteTool
NewUserAPI Temporal reference UserAPI
ImprovedParser References history Parser
ToolFactory Pattern name noise Tool or createTool()
AbstractToolInterface Redundant qualifiers Tool
executeToolWithValidation() Implementation in name execute()

When Changing Code

Rule: Never document old behavior or the change in names.

Red Flags - STOP and Rename

If you catch yourself writing:

  • "New", "Old", "Legacy", "Improved", "Enhanced"
  • "Unified", "Refactored", "Updated", "Modern"
  • Implementation details ("Zod", "JSON", "MCP", "SQL")
  • Unnecessary pattern names ("Factory", "Builder", "Manager")
  • Redundant qualifiers ("Abstract", "Base", "Interface")

STOP. Find a name describing actual purpose in the domain.

Common Rationalizations

Excuse Reality
"Need to distinguish from old version" Old version shouldn't exist or should be in different namespace.
"New developers need to know it's improved" Code quality shows in behavior, not names.
"Factory pattern is important here" If pattern is core purpose, fine. Usually it's not.
"Everyone knows what Zod is" Today they do. Names should outlive dependencies.
"It IS a wrapper around MCP" That's implementation. What does it DO in your domain?

Verification

Before committing names:

  • Name describes domain purpose
  • No implementation details
  • No temporal context
  • No unnecessary pattern names
  • Forms readable sentences with other code
  • No "new", "old", "improved", "wrapper"

Real-World Examples

Bad Naming (Don't Do This)

typescript
class ImprovedZodConfigValidator { }           // ❌ Temporal + implementation
const newAPIClientWithRetry = new Client();    // ❌ Temporal + implementation
function executeEnhancedToolFactory() { }      // ❌ Temporal + pattern noise

// Using them
const validator = new ImprovedZodConfigValidator();
validator.validateWithNewSchema();

Good Naming

typescript
class ConfigValidator { }                      // ✅ Domain purpose
const apiClient = new Client();               // ✅ What it is
function createTool() { }                     // ✅ What it does

// Using them - reads like domain language
const validator = new ConfigValidator();
validator.validate();

Integration with Other Skills

For tactical variable naming: See skills/naming-variables for comprehensive variable naming techniques (optimal length, scope rules, conventions for booleans/collections/qualifiers, naming as diagnostic tool)

For comment guidelines: See skills/writing-evergreen-comments for keeping comments evergreen (no temporal context in comments either)

Didn't find tool you were looking for?

Be as detailed as possible for better results