vercel-labs/zerolang — reverse-engineered prompt

Reverse engineered prompt

GitHub

Build me an experimental programming language called Zero that is made for AI coding agents, not mainly for humans.

I want a working command line tool where someone can check a Zero file, run simple programs, build an executable, format or inspect code, and get structured output that an agent can understand and use to fix problems. Include small examples like hello world and basic math, plus clear docs that explain the language by example instead of assuming people will memorize syntax.

Keep the language small, regular, and predictable. The standard library should cover common things so users don’t immediately need to hunt for dependencies. Add JSON style diagnostics, graph info, size reports, and a doctor command so agents can debug issues.

Please include tests, conformance checks, benchmarks if useful, and a quick start guide. Make it obvious this is experimental, unstable, and not safe for production or sensitive data. Look up current docs online if you need to.

Want more depth? Deep Reverse