daletoniris/tokioai ? reverse-engineered prompt

Build me a Python terminal app called TokioAI that acts like an autonomous AI assistant for cybersecurity, DevOps, engineering, and general system tasks.

I want it to start in an interactive shell, let me type normal questions, and also support single command prompts from the terminal. It should be able to switch between multiple AI providers at runtime, including Claude, OpenAI, Gemini, OpenRouter, and Ollama, and ask for setup the first time so it can save credentials locally without me editing files by hand.

Please include useful slash commands for status, health, threats, logs, WiFi, Home Assistant, robots, and other live system checks, plus model switching, tab completion, multi line input, cost tracking, and session saving. It should feel polished, handle Ctrl+C cleanly, mask secrets, and work on Linux, macOS, Windows, and WSL. If you need current docs for provider setup or APIs, look them up online.