本文へ移動
cccskills
無料GitHub で公開

environment-config-manager

Environment configuration manager for 12-factor config, feature flags, and multi-environment management. Activate on: environment variables, ConfigMap, feature flags, 12-factor config, dotenv management, config per environment, runtime config. NOT for: secret storage (use secret-management-expert), IaC provisioning (use terraform-module-builder), CI/CD variables (use github-actions-pipeline-builder).

インストール方法を見る

含まれるファイル(1)

  • SKILL.md5.8 KB

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。

Environment Config Manager

Expert in 12-factor configuration management, feature flags, and environment-specific config across dev, staging, and production.

Activation Triggers

Activate on: "environment variables", "ConfigMap", "feature flags", "12-factor config", "dotenv setup", "config per environment", "runtime configuration", "config validation", "environment parity"

NOT for: Secret storage → secret-management-expert | IaC provisioning → terraform-module-builder | CI/CD variables → github-actions-pipeline-builder

Quick Start

  1. Separate config from code — no hardcoded URLs, ports, or feature toggles in source
  2. Define config schema — typed validation with defaults (zod, envalid, convict)
  3. Layer environments — base config overridden by environment-specific values
  4. Implement feature flags — runtime toggleable features without deployment
  5. Validate on startup — fail fast if required config is missing or invalid

Core Capabilities

DomainTechnologies
Validationzod, envalid, @t3-oss/env-nextjs, convict, joi
Feature FlagsLaunchDarkly, Unleash, Flagsmith, Statsig, simple JSON flags
K8s ConfigConfigMap, ExternalSecret, Kustomize overlays
Local Devdotenv, direnv, 1Password CLI (op run), docker-compose env
Runtime Configetcd, Consul KV, AWS AppConfig, Firebase Remote Config

Architecture Patterns

Typed Config Validation (TypeScript)

// config/env.ts — fail fast on invalid config
import { z } from 'zod';

const envSchema = z.object({
  NODE_ENV: z.enum(['development', 'staging', 'production']),
  PORT: z.coerce.number().default(3000),
  DATABASE_URL: z.string().url(),
  REDIS_URL: z.string().url().optional(),
  LOG_LEVEL: z.enum(['debug', 'info', 'warn', 'error']).default('info'),
  FEATURE_NEW_CHECKOUT: z.coerce.boolean().default(false),
  API_RATE_LIMIT: z.coerce.number().default(100),
});

export type Env = z.infer<typeof envSchema>;

// Parse and validate — throws on startup if invalid
export const env = envSchema.parse(process.env);

Environment Layering

.env                    # Base defaults (committed, no secrets)
.env.development        # Dev overrides (committed)
.env.staging            # Staging overrides (committed, no secrets)
.env.production         # Production overrides (committed, no secrets)
.env.local              # Local overrides (gitignored, may have secrets)

Loading order (later overrides earlier):
  .env → .env.{NODE_ENV} → .env.local

Secrets come from secret store, NOT .env files:
  DATABASE_URL → AWS Secrets Manager / Vault
  API_KEY → External Secrets Operator

Feature Flag Architecture

┌──────────────┐    ┌──────────────┐    ┌──────────────┐
│   Client     │    │  Flag Store   │    │   Admin UI   │
│  SDK/hook    │───▶│  (Unleash/   │◀───│  Toggle flags │
│              │    │   LaunchDarkly)│    │  per env     │
└──────────────┘    └──────────────┘    └──────────────┘

Evaluation: user context + flag rules → boolean/variant
  - Percentage rollout: 10% of users see new feature
  - User targeting: beta users see new feature
  - Environment: enabled in staging, disabled in production
  - Kill switch: instantly disable without deployment

Anti-Patterns

  1. Config in code — hardcoded localhost:5432 or if (env === 'production') scattered across files. Centralize all config in one validated module.
  2. Secrets in .env files committed to Git — even example files leak secret formats. Use .env.example with placeholder descriptions, not values.
  3. No config validation — trusting process.env.PORT is a valid number. Validate and coerce all config at startup with a schema library.
  4. Feature flags as permanent config — flags accumulate and become tech debt. Set expiration dates and clean up old flags quarterly.
  5. Environment-specific code paths — if (process.env.NODE_ENV === 'production') for business logic. Use feature flags instead; environment should only affect infrastructure config.

Quality Checklist

[ ] All config loaded from environment variables (12-factor compliant)
[ ] Config schema validated on application startup (fail fast)
[ ] .env.example documents all required variables with descriptions
[ ] No secrets in committed .env files
[ ] .env.local in .gitignore
[ ] Feature flags have owner, description, and expiration date
[ ] Dev/staging/production parity maintained (same config keys)
[ ] Config changes do not require code deployment
[ ] Default values sensible for development (zero-config local setup)
[ ] ConfigMaps and Secrets separated in Kubernetes
[ ] Feature flag cleanup tracked in backlog
[ ] Runtime config changes logged and auditable

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

Expert in 2000s-era music visualization (Milkdrop, AVS, Geiss) and modern WebGL implementations. Specializes in Butterchurn integration, Web Audio API AnalyserNode FFT data, GLSL shaders for audio-reactive visuals, and psychedelic generative art. Activate on "Milkdrop", "music visualization", "WebGL visualizer", "Butterchurn", "audio reactive", "FFT visualization", "spectrum analyzer". NOT for simple bar charts/waveforms (use basic canvas), video editing, or non-audio visuals.

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。

curiositech/port-daddy22026年10月8日 更新

Expert legal research agent for finding and scraping expungement data state by state. Knows authoritative sources, URL patterns, Firecrawl configuration, and 2026 legal landscape.

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。

curiositech/port-daddy22026年10月8日 更新

Expert in 3D computer vision labeling tools, workflows, and AI-assisted annotation for LiDAR, point clouds, and sensor fusion. Covers SAM4D/Point-SAM, human-in-the-loop architectures, and vertical-specific training strategies. Activate on '3D labeling', 'point cloud annotation', 'LiDAR labeling', 'SAM 3D', 'SAM4D', 'sensor fusion annotation', '3D bounding box', 'semantic segmentation point cloud'. NOT for 2D image labeling (use clip-aware-embeddings), general ML training (use ml-engineer), video annotation without 3D (use computer-vision-pipeline), or VLM prompt engineering (use prompt-engineer).

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。

curiositech/port-daddy22026年10月8日 更新

Apply crisis decision-making research to agent routing, uncertainty triage, and coordination failure analysis in time-pressured systems. Use when diagnosing handoff failures, analytical paralysis, or expert judgment under incomplete information. NOT for routine coding, simple CRUD design, or static single-agent tasks with complete information.

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。

curiositech/port-daddy22026年10月8日 更新

Use for insight, reframing, contradiction, impasse, and anomaly-driven problem solving when execution effort no longer helps. NOT for routine optimization, error correction, or well-specified tasks with known solution paths.

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。

curiositech/port-daddy22026年10月8日 更新

Apply cognitive task analysis to expert work that depends on perceptual cues, branching judgment, and recurring monitoring loops. Use when decomposing expert capability into agent structure, simulation design, or validation interviews. NOT for ordinary step-by-step SOP capture or simple pipelines with no tacit cue layer.

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。

curiositech/port-daddy22026年10月8日 更新

curiositech のスキルをすべて見る

このスキルの問題を報告する