krishnashahane/Genesis ? reverse-engineered prompt

Build me an open source runtime for autonomous AI agents that feels like a tiny operating system for them.

I want to be able to give it a goal from the command line or through a web app, and have it run a full loop like plan, research, build, review, test, debug, optimize, reflect, and store what it learned so it gets better next time. It should work completely offline by default with no API keys needed, using a mock AI model and in memory storage, but it should also be able to plug into real providers and optional Redis, PostgreSQL, and vector memory later.

Please include a simple REST API, a live web interface, event driven logging, task queue support, agent memory, and a way to inspect agents, tools, events, and metrics. Make it clean, testable, and easy to run with one command, and look up current docs online if you need to.