Fix silent response loss and streaming race condition

Add streaming, queue, /help, /project, and token display

Improve bot UX: formatting, reply context, and cancel support

Improve image handling with GPT-4o vision and Claude integration

Fix critical memory issues in long-running bot process

Add inactivity detection with Y/N prompt and session context

Replace Ollama vision with OpenAI gpt-4o for image analysis

Replace Anthropic SDK with local Ollama for image analysis

Add image analysis support via Claude Vision API

Render markdown links and formatting in Telegram replies

Add GitHub token auth and improve status UI

Initial commit: Telegram Claude Code bot bridge