Hainrixz/cyber-neo — reverse-engineered prompt

Build me an open source cybersecurity agent that runs as a Claude Code skill called Cyber Neo. I want to be able to type /cyber neo with a project path, or just scan the current folder, and have it audit a codebase for common app security problems without needing security expertise. It should check code issues, auth and access control, crypto mistakes, exposed secrets, dependency risks, web security problems, supply chain concerns, workflow security, Docker and container issues, error handling, and logging leaks.

Make it work especially well for JavaScript, TypeScript, and Python projects, but still catch generic risky patterns in any language. If optional security tools are available, use them, otherwise still do a solid built in analysis. I want the scan to feel thorough and fast, with parallel analysis where it makes sense.

The output should be a clean professional markdown report saved to the Desktop with severity, OWASP 2025 and CWE mapping, clear priority order, file locations, and simple fix guidance with examples. Look up current Claude Code skill docs online if you need to.