Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Set up environment variables, .env files, and configuration management. Use when configuring environment variables, creating .env files, or managing app configuration.
.claude/skills/onewave-ai-env-setup-wizard/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-07 | ✗→✓ | ▲ Improved | 35% | 0% |
| case-01 | ✓→✓ | = Same ✓ | 88% | 0% |
| case-02 | ✓→✓ | = Same ✓ | 138% | 0% |
| case-03 | ✓→✓ | = Same ✓ | 27% | 0% |
| case-04 | ✓→✓ | = Same ✓ | 38% | 0% |
When setting up environment configuration:
project/
├── .env # Local development (git-ignored)
├── .env.example # Template (committed to git)
├── .env.local # Local overrides (git-ignored)
├── .env.development # Development defaults
├── .env.production # Production defaults
└── src/
└── lib/
└── env.ts # Type-safe env accessbash# =================== # Application # =================== NODE_ENV=development APP_URL=http://localhost:3000 PORT=3000 # =================== # Database # =================== DATABASE_URL=postgresql://user:password@localhost:5432/dbname # =================== # Authentication # =================== # Generate with: openssl rand -base64 32 JWT_SECRET= NEXTAUTH_SECRET= NEXTAUTH_URL=http://localhost:3000 # =================== # Third-party APIs # =================== # Get from: https://stripe.com/dashboard STRIPE_SECRET_KEY= STRIPE_PUBLISHABLE_KEY= STRIPE_WEBHOOK_SECRET= # Get from: https://resend.com RESEND_API_KEY= # =================== # Storage # =================== AWS_ACCESS_KEY_ID= AWS_SECRET_ACCESS_KEY= AWS_REGION=us-east-1 S3_BUCKET_NAME=
typescript// src/lib/env.ts import { z } from 'zod'; const envSchema = z.object({ // App NODE_ENV: z.enum(['development', 'production', 'test']).default('development'), APP_URL: z.string().url(), PORT: z.coerce.number().default(3000), // Database DATABASE_URL: z.string().min(1), // Auth JWT_SECRET: z.string().min(32), NEXTAUTH_SECRET: z.string().min(32), NEXTAUTH_URL: z.string().url(), // APIs (optional in dev) STRIPE_SECRET_KEY: z.string().optional(), RESEND_API_KEY: z.string().optional(), }); // Validate on import const parsed = envSchema.safeParse(process.env); if (!parsed.success) { console.error('Invalid environment variables:'); console.error(parsed.error.flatten().fieldErrors); process.exit(1); } export const env = parsed.data; // Type export for use elsewhere export type Env = z.infer<typeof envSchema>;
typescript// src/lib/env.ts for Next.js import { z } from 'zod'; // Server-side variables const serverSchema = z.object({ DATABASE_URL: z.string(), JWT_SECRET: z.string(), }); // Client-side variables (must start with NEXT_PUBLIC_) const clientSchema = z.object({ NEXT_PUBLIC_APP_URL: z.string().url(), NEXT_PUBLIC_STRIPE_KEY: z.string(), }); export const serverEnv = serverSchema.parse(process.env); export const clientEnv = clientSchema.parse({ NEXT_PUBLIC_APP_URL: process.env.NEXT_PUBLIC_APP_URL, NEXT_PUBLIC_STRIPE_KEY: process.env.NEXT_PUBLIC_STRIPE_KEY, });
typescript// src/env.mjs import { createEnv } from "@t3-oss/env-nextjs"; import { z } from "zod"; export const env = createEnv({ server: { DATABASE_URL: z.string().url(), NODE_ENV: z.enum(["development", "test", "production"]), }, client: { NEXT_PUBLIC_APP_URL: z.string().url(), }, runtimeEnv: { DATABASE_URL: process.env.DATABASE_URL, NODE_ENV: process.env.NODE_ENV, NEXT_PUBLIC_APP_URL: process.env.NEXT_PUBLIC_APP_URL, }, });
gitignore# Environment files .env .env.local .env.*.local # Keep example !.env.example
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-10 | fail→fail | 12,151 | 11,788 | -3% | 1 | 1 | 0% | 2,373 | 3,431 | +45% | 0 | 0 | — |
case-01 | pass→pass | 6,098 | 35,381 | +480% | 1 | 1 | 0% | 1,162 | 2,184 | +88% | 0 | 0 | — |
case-02 | pass→pass | 5,354 | 6,052 | +13% | 1 | 1 | 0% | 951 | 2,264 | +138% | 0 | 0 | — |
case-03 | pass→pass | 31,669 | 8,411 | -73% | 1 | 1 | 0% | 2,125 | 2,703 | +27% | 0 | 0 | — |
case-04 | pass→pass | 13,262 | 10,778 | -19% | 1 | 1 | 0% | 2,032 | 2,794 | +38% | 0 | 0 | — |
case-05 | pass→pass | 5,812 | 6,508 | +12% | 1 | 1 | 0% | 720 | 2,271 | +215% | 0 | 0 | — |
case-06 | pass→pass | 6,284 | 4,914 | -22% | 1 | 1 | 0% | 967 | 1,994 | +106% | 0 | 0 | — |
case-07 | fail→pass | 8,342 | 3,990 | -52% | 1 | 1 | 0% | 1,394 | 1,882 | +35% | 0 | 0 | — |
case-08 | pass→pass | 2,812 | 2,950 | +5% | 1 | 1 | 0% | 378 | 1,715 | +354% | 0 | 0 | — |
case-09 | pass→pass | 12,546 | 12,694 | +1% | 1 | 1 | 0% | 2,255 | 3,171 | +41% | 0 | 0 | — |
case-11 | pass→pass | 15,369 | 12,164 | -21% | 1 | 1 | 0% | 2,693 | 3,356 | +25% | 0 | 0 | — |
case-12 | pass→pass | 2,970 | 2,647 | -11% | 1 | 1 | 0% | 424 | 1,649 | +289% | 0 | 0 | — |
case-13 | pass→pass | 6,996 | 7,434 | +6% | 1 | 1 | 0% | 1,131 | 2,543 | +125% | 0 | 0 | — |
case-14 | pass→pass | 2,380 | 1,529 | -36% | 1 | 1 | 0% | 314 | 1,458 | +364% | 0 | 0 | — |
case-15 | pass→pass | 3,498 | 3,312 | -5% | 1 | 1 | 0% | 567 | 1,851 | +226% | 0 | 0 | — |
case-16 | pass→pass | 11,574 | 7,455 | -36% | 1 | 1 | 0% | 2,298 | 2,541 | +11% | 0 | 0 | — |
case-17 | pass→pass | 5,830 | 3,056 | -48% | 1 | 1 | 0% | 1,140 | 1,770 | +55% | 0 | 0 | — |
case-18 | fail→fail | 13,781 | 10,981 | -20% | 1 | 1 | 0% | 2,516 | 3,275 | +30% | 0 | 0 | — |
case-19 | pass→pass | 6,864 | 5,368 | -22% | 1 | 1 | 0% | 1,273 | 2,150 | +69% | 0 | 0 | — |
case-20 | pass→pass | 7,550 | 7,951 | +5% | 1 | 1 | 0% | 1,457 | 2,639 | +81% | 0 | 0 | — |
case-21 | pass→pass | 4,868 | 5,088 | +5% | 1 | 1 | 0% | 758 | 2,097 | +177% | 0 | 0 | — |
case-22 | pass→pass | 3,420 | 3,716 | +9% | 1 | 1 | 0% | 600 | 1,856 | +209% | 0 | 0 | — |
case-23 | pass→pass | 3,793 | 4,194 | +11% | 1 | 1 | 0% | 606 | 1,963 | +224% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 23 cases were attempted. The headline lift of +4 percentage points is the difference between those two pass rates over the 23 comparable cases.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.