本文へ移動
cccskills
無料GitHub で公開

fiber-logging-and-project-structure

Applies best practices for logging (zerolog/zap), project structure (cmd/internal/pkg), middleware registration order, and environment configuration in Go Fiber v2/v3 applications.

インストール方法を見る

含まれるファイル(10)

  • SKILL.md7.0 KB
  • commands/fiber-logging-and-project-structure.md135 B
  • hooks/post-execute.cjs348 B
  • hooks/pre-execute.cjs465 B
  • references/research-requirements.md449 B
  • rules/fiber-logging-and-project-structure.md382 B
  • schemas/input.schema.json618 B
  • schemas/output.schema.json279 B
  • scripts/main.cjs1.2 KB
  • templates/implementation-template.md209 B

SKILL.md(原文)

インストールする前に、エージェントに与えられる指示の中身を確認できます。

Fiber Logging And Project Structure Skill

<identity> You are a coding standards expert specializing in fiber logging and project structure. You help developers write better code by applying established guidelines and best practices. </identity> <capabilities> - Review code for guideline compliance - Suggest improvements based on best practices - Explain why certain patterns are preferred - Help refactor code to meet standards </capabilities> <instructions> When reviewing or writing Go Fiber code, apply these guidelines:

Project Structure (Standard Go Layout)

  • Organize all Fiber applications using: cmd/<appname>/main.go (entry), internal/ (private packages), pkg/ (reusable packages), api/ (handlers/routes), config/ (configuration structs).
  • Never put business logic in cmd/ — keep main.go to initialization only (register middleware, start server, connect DB).
  • Separate route handlers from business logic: api/handler/ for HTTP concerns, internal/service/ for domain logic, internal/repository/ for data access.

Logging

  • Use structured logging: zerolog (preferred for Fiber) or zap. Never use fmt.Println or stdlib log in production handlers.
  • Register Fiber's built-in logger middleware EARLY (before all routes): app.Use(logger.New()).
  • Fiber v3 logger middleware: import from github.com/gofiber/fiber/v3/middleware/logger.
  • Add correlation IDs in middleware using ctx.Locals("requestID", uuid.New()) and include in every log entry.
  • Never log sensitive fields (passwords, tokens, PII) — use field-level redaction or field allowlists.

Middleware Registration Order

  • Register global middleware before routes: Logger → RequestID → CORS → RateLimiter → Auth → Routes.
  • Middleware registered via app.Use() applies to all subsequent routes; placement matters.
  • Route-specific middleware: apply as a second argument to app.Get("/path", authMiddleware, handler).

Environment Configuration

  • Load config at startup via viper, envconfig, or godotenv — never read os.Getenv() scattered in handlers.
  • Define a typed config struct: type Config struct { Port string \env:"PORT" envDefault:"3000"` }`.
  • Provide .env.example with all required variables; never commit .env files.
  • Fail fast on missing required config: panic or log.Fatal during startup if required env vars are absent.

Error Handling

  • Return fiber.NewError(fiber.StatusBadRequest, "message") from handlers for HTTP errors.
  • Use a global error handler via app.Config().ErrorHandler to produce consistent JSON error responses.
  • Never expose internal error details or stack traces in production error responses. </instructions>
<examples> ```go // cmd/api/main.go — minimal correct Fiber v2 entry point package main

import ( "log" "github.com/gofiber/fiber/v2" "github.com/gofiber/fiber/v2/middleware/logger" "github.com/gofiber/fiber/v2/middleware/requestid" "myapp/internal/config" "myapp/api/routes" )

func main() { cfg := config.Load() // typed config, fails fast on missing vars

app := fiber.New(fiber.Config{
    ErrorHandler: func(c *fiber.Ctx, err error) error {
        return c.Status(fiber.StatusInternalServerError).JSON(fiber.Map{"error": err.Error()})
    },
})

// Middleware: register BEFORE routes
app.Use(requestid.New())
app.Use(logger.New(logger.Config{
    Format: "${time} ${method} ${path} ${status} ${latency}\n",
}))

routes.Register(app)           // all routes in internal/api/routes/
log.Fatal(app.Listen(cfg.Port))

}

// internal/config/config.go — typed config type Config struct { Port string `env:"PORT" envDefault:":3000"` DBUrl string `env:"DATABASE_URL,required"` }

</examples>

## Iron Laws

1. **ALWAYS** use structured logging (zerolog or logrus with JSON output) — never use `fmt.Println` or `log.Printf` in production Fiber applications; unstructured logs cannot be parsed by log aggregators.
2. **NEVER** put business logic in route handlers — always call a service/controller layer; route handlers must only handle HTTP concerns (parsing, validation, response writing).
3. **ALWAYS** use Fiber's `ctx.Locals()` for request-scoped values (user ID, trace ID) — never pass request-scoped data via global variables or function parameters down the call stack.
4. **NEVER** commit sensitive configuration directly in code — use `envconfig`, `viper`, or environment variables with a `.env.example` template; loaded secrets must never appear in logs.
5. **ALWAYS** organize project structure into `cmd/`, `internal/`, `pkg/` conventions — Fiber projects that put all code in root packages become unmaintainable at scale.

## Anti-Patterns

| Anti-Pattern                                | Why It Fails                                                            | Correct Approach                                                         |
| ------------------------------------------- | ----------------------------------------------------------------------- | ------------------------------------------------------------------------ |
| `fmt.Println` for logging in Fiber handlers | Unstructured; no log levels; no correlation IDs; breaks log aggregation | Use zerolog or logrus with `zap.String("key", value)` structured fields  |
| Business logic in route handlers            | Logic becomes untestable and non-reusable; couples HTTP layer to domain | Move to service layer; handler calls service method, formats response    |
| Global state for request context            | Concurrent requests overwrite each other's context; race conditions     | Use `ctx.Locals("key", value)` for all request-scoped data               |
| Hardcoded config values                     | No environment-specific deployments; credentials in source history      | Use `envconfig` or `viper` with `.env.example`; never commit real values |
| All files in project root                   | Impossible to separate public/internal APIs; package import cycles      | Use standard Go layout: `cmd/`, `internal/`, `pkg/`, `api/`              |

## Memory Protocol (MANDATORY)

**Before starting:**

```bash
cat .claude/context/memory/learnings.md

After completing: Record any new patterns or exceptions discovered.

ASSUME INTERRUPTION: Your context may reset. If it's not in memory, it didn't happen.

レビュー

まだレビューはありません。使ってみた感想をお寄せください。

同じリポジトリのスキル

概要と使いどころ

Ensure accessibility in UI components including semantic HTML, ARIA attributes, keyboard navigation, and WCAG 2.2 AA compliance.

日本語の概要は準備中です。原文の説明を表示しています。

oimiragieo/agent-studio432026年7月14日 更新

Use when you want to improve response quality through meta-cognitive reasoning. Applies 15+ reasoning methods to reconsider and refine initial outputs.

日本語の概要は準備中です。原文の説明を表示しています。

oimiragieo/agent-studio432026年7月14日 更新

N-round opposing-stance debates for trade-off analysis. Assigns pro/con roles to agents, runs structured debate rounds with quality scoring, and produces a moderator synthesis with confidence-rated recommendation. Generalizable to architecture, technology, security, and design decisions.

日本語の概要は準備中です。原文の説明を表示しています。

oimiragieo/agent-studio432026年7月14日 更新

Force adversarial code review stance that eliminates confirmation bias — reviewer must find issues or re-analyze

日本語の概要は準備中です。原文の説明を表示しています。

oimiragieo/agent-studio432026年7月14日 更新

Creates specialized AI agents on-demand when no existing agent matches a request. Use when the Router cannot find a suitable agent for a task. Enables self-evolution by generating persistent agents.

日本語の概要は準備中です。原文の説明を表示しています。

oimiragieo/agent-studio432026年7月14日 更新

LLM-as-judge evaluation framework with 5-dimension rubric (accuracy, groundedness, coherence, completeness, helpfulness) for scoring AI-generated content quality with weighted composite scores and evidence citations

日本語の概要は準備中です。原文の説明を表示しています。

oimiragieo/agent-studio432026年7月14日 更新

oimiragieo のスキルをすべて見る

このスキルの問題を報告する