Golang Best Practices (Meta-Skill)

Comprehensive Go code review skill coordinating 5 specialized domain skills for complete code audits.

[!NOTE] v2.0.0 Update: This skill now coordinates multiple focused domain skills. Use specific skills for targeted reviews or this meta-skill for comprehensive audits.

Available Skills

1. Concurrency Safety ✅

12 rules | Goroutines, channels, race conditions, deadlocks
Status: Active (v2.0.0)

Use when:

• Reviewing code with goroutines or channels
• Debugging race conditions or deadlocks
• Auditing concurrent data access

Trigger phrases: "Check for race conditions", "Review concurrency", "Find goroutine leaks"

2. Clean Architecture ✅

9 rules | Layer separation, dependency rules, gRPC patterns
Status: Active (v2.0.0)

Use when:

• Auditing service architecture
• Reviewing layered architecture compliance
• Ensuring proper dependency injection

Trigger phrases: "Audit architecture", "Check layer dependencies", "Review Clean Architecture"

3. Error Handling ✅

7 rules | Error wrapping, context, nil checks
Status: Active (v2.0.0)

Use when:

• Reviewing error propagation
• Checking context usage
• Auditing error handling patterns

Trigger phrases: "Review error handling", "Check error wrapping", "Verify context propagation"

4. Design Patterns ✅

13 rules | Code smells, refactoring, Gang of Four patterns
Status: Active (v2.0.0)

Use when:

• Refactoring complex code
• Reducing technical debt
• Applying design patterns

Trigger phrases: "Refactor this code", "Reduce complexity", "Apply design patterns"

5. Idiomatic Go ✅

6 rules | Go-specific idioms, interfaces, pointers
Status: Active (v2.0.0)

Use when:

• Ensuring idiomatic Go style
• Reviewing interface usage
• Optimizing pointer usage

Trigger phrases: "Is this idiomatic Go?", "Review Go style", "Check interface design"

When to Use This Meta-Skill

Use this meta-skill for:

• Complete codebase audits
• Pre-production reviews
• Unknown issue categories
• Comprehensive refactoring

For targeted reviews, use specific skills above (e.g., just concurrency-safety or clean-architecture).

Rule Categories by Priority

critical-

high-

medium-

arch-

Quick Reference

1. Critical Issues (CRITICAL)

These prevent bugs, crashes, and production failures

• critical-error-wrapping

  - Use %w for error wrapping, not %v

• critical-defer-in-loop

  - Avoid defer in loops (resource leaks)

• critical-context-leak

  - Always defer cancel() after context creation

• critical-error-shadow

  - Don't shadow err variable in nested scopes

• critical-goroutine-leak

  - Goroutines must have exit conditions

• critical-race-condition

  - Protect shared state with mutex/channels

• critical-channel-deadlock

  - Ensure paired send/receive operations

• critical-close-panic

  - Sender closes channel, not receiver

2. High-Impact Patterns (HIGH)

Reliability and correctness improvements

• high-pointer-receiver

  - Use pointer receivers for mutations

• high-context-propagation

  - Propagate context through call chain

• high-error-is-as

  - Use errors.Is/As instead of ==

• high-interface-nil

  - Check interface nil correctly

• high-goroutine-unbounded

  - Limit concurrent goroutines (worker pool)

• high-channel-not-closed

  - Always close channels when done

• high-loop-variable-capture

  - Avoid closure over loop variables

• high-waitgroup-mismatch

  - Match Add() and Done() calls

• high-business-logic-handler

  - Keep delivery layer thin

• high-business-logic-repository

  - No business logic in data layer

• high-constructor-creates-deps

  - Inject dependencies, don't create

• high-transaction-in-repository

  - Transactions belong in usecase

• high-god-object

  - Extract logic from 300+ line functions

• high-extract-method

  - Name complex code blocks with descriptive methods

3. Medium Improvements (MEDIUM)

Code quality and idioms

• medium-interface-pollution

  - Keep interfaces small (<5 methods)

• medium-accept-interface-return-struct

  - API flexibility pattern

• medium-pointer-overuse

  - Don't overuse pointers for small types

• medium-directional-channels

  - Use send/receive-only channels

• medium-buffered-channel-size

  - Choose appropriate buffer size

• medium-select-default

  - Avoid busy-wait with select

• medium-fat-interface

  - Split large interfaces

• medium-usecase-complexity

  - Move business logic to domain entities

• medium-interface-in-implementation

  - Define interfaces where used

• medium-sentinel-error-usage

  - Use sentinel errors for stable categories

• medium-primitive-obsession

  - Replace primitives with value objects

• medium-long-parameter-list

  - Use parameter objects for >5 params

• medium-data-clumps

  - Extract repeated parameter groups

• medium-feature-envy

  - Move logic closer to data

• medium-magic-constants

  - Replace magic numbers with named constants

• medium-builder-pattern

  - Fluent API for complex construction

• medium-factory-constructor

  - Validated object creation

• medium-introduce-parameter-object

  - Group related parameters

• medium-switch-to-strategy

  - Replace type switches with interfaces

• medium-middleware-decorator

  - Decorator pattern for http.Handler

• medium-law-of-demeter

  - Reduce coupling, avoid message chains

4. Architecture (ARCHITECTURE)

Clean Architecture compliance for gRPC/usecase/repository pattern

• arch-domain-import-infra

  - Domain must not import infrastructure

• arch-concrete-dependency

  - Depend on interfaces, not concrete types

• arch-repository-business-logic

  - Repositories do CRUD only

• arch-usecase-orchestration

  - Usecases orchestrate, entities decide

• arch-interface-segregation

  - Small, consumer-defined interfaces

How to Use

For Code Review

1. Read the code file(s)
2. Check against rules in priority order (Critical first)
3. For each violation found, reference the specific rule file
4. Provide exact line numbers and explanation
5. Show corrected code example

For Refactoring

1. Identify code smells matching detection patterns
2. Apply fixes from corresponding rule files
3. Verify no regressions introduced
4. Run tests to confirm correctness

Accessing Detailed Rules

Each rule file contains:

• Brief explanation of why it matters
• Detection criteria (how to spot the issue)
• Incorrect code example with explanation
• Correct code example with explanation
• Impact assessment
• Additional context and references

Example:

rules/critical-error-wrapping.md
rules/high-pointer-receiver.md

Reference Guides

For deep dives on specific topics:

• references/concurrency-deep-dive.md

  - Comprehensive concurrency patterns

• references/error-handling-guide.md

  - Complete error handling strategies

• references/testing-strategies.md

  - Go testing best practices

Common Usage Patterns

Review entire file:

Review this Go file for anti-patterns and suggest improvements

Focus on specific category:

Check this code for concurrency issues (CRITICAL level)

Architecture audit:

Verify this service follows Clean Architecture principles

Performance optimization:

Find performance issues in this Go code

Trigger Phrases

For comprehensive audits (uses this meta-skill):

• "Review my Go code"
• "Check this Golang file"
• "Find Go anti-patterns"
• "Audit this Go service"

For domain-specific reviews (uses specific skills):

• "Check for race conditions" → concurrency-safety
• "Audit architecture" → clean-architecture
• "Review error handling" → error-handling
• "Refactor this code" → design-patterns
• "Is this idiomatic Go?" → idiomatic-go

Best Practices Philosophy

Based on research from authoritative sources:

Jon Bodner's Principles:

• Go is deliberately simple and explicit
• "Boring" code is good code - predictable and maintainable
• Understand the "why" behind language features
• Idiomatic Go prioritizes clarity over cleverness

Katherine Cox-Buday's Guidelines:

• Concurrency is not parallelism
• Channels are for communication, mutexes are for state
• Always have exit conditions for goroutines
• Context is the standard way to propagate cancellation

Clean Architecture (Uncle Bob):

• Dependencies point inward toward business logic
• Domain layer has no external dependencies
• Interfaces defined by consumers, not producers
• Separation of concerns across layers

Output Format

When reviewing code, use this format:

## Critical Issues Found: X

### [Rule Name] (Line Y)
**Issue**: Brief description
**Fix**: Suggested correction
**Example**:
```go
// Corrected code here

Medium Improvements: X

[Similar format for medium priority items]

## Notes

- Rules are evidence-based from authoritative Go books
- Detection patterns help identify issues programmatically
- All rules include working code examples
- Tailored for gRPC/usecase/repository architecture patterns