Multi-Environment Isolator

This skill appears to implement the described multi-environment scaffolding, but take these precautions before running it on a real project: 1. Back up the target project directory first (or try it in a copy/sandbox). The scripts will create files and update .gitignore. 2. Documentation mismatch: SKILL.md mixes examples for two script versions. If you want frontend flags like --dev-backend-port/--frontend-dir, run scripts/setup_envs_v2.py; the older scripts/setup_envs.py accepts different flag names. Verify which script you run and its accepted flags. 3. Immediately change generated secrets: replace JWT_SECRET (and any other secrets) in .env.prod with a strong random value before using a production environment. The defaults are intentionally weak placeholders. 4. Review package.json and dependencies before allowing the generated frontend start scripts to run. The frontend start scripts will run `npm install` (if node_modules missing) and use `npx playwright` for tests — npm installs run arbitrary install/postinstall scripts and fetch packages from the network, so only run them in a trusted environment or after checking dependencies. 5. Verify ports and database choices: defaults differ between versions; ensure ports don't conflict on your machine and set DATABASE_URL for production if you need a managed DB instead of local sqlite. 6. If you only need file generation (no npm/network activity), inspect and run the setup script but avoid starting frontend scripts until you've reviewed the project and dependencies. Overall: the code is not obviously malicious, but the doc/version mismatch and the automated npm behavior justify caution. If you want, I can point out the exact command flags each script accepts or produce a safe checklist of what to inspect in package.json before running npm install.