jcarlos78/GAIDE ? reverse-engineered prompt

Build me a project template for AI assisted software development that feels safe and disciplined, not like random vibe coding.

I want the repo to come ready for Claude Code, Cursor, Windsurf, and Codex, with a clear workflow that starts from a spec, then a plan, then tasks, then implementation, with human approval at each step. Set up a simple folder structure for specs, docs, tests, and source code, plus a project briefing file the agent can auto read.

Also include reusable agent instructions for writing specs, reviewing code, reviewing security, generating tests, and verifying features end to end. Add basic guardrails so secrets are blocked, obvious security issues are checked, edited files get validated, and a turn cannot end with broken tests or leaked secrets. I also want a starter example feature, some documentation for how to use the workflow, and a bootstrap script that checks the project is healthy before work starts. If you need to, look up current docs online for any tool integrations.