Add eye2byte skill — screen context for coding agents via MCP

[wolverin0]

Summary

• Adds eye2byte skill under skills/wolverin0/eye2byte/
• Eye2byte gives any MCP-compatible agent the ability to see the user's actual screen
• Captures screenshots, voice narration, and annotations from any monitor or device
• Feeds captures to a vision model and returns structured Context Packs

What Eye2byte does

Tool	Description
capture_and_summarize	Screenshot + vision analysis with multi-monitor support
capture_with_voice	Screenshot + voice recording + transcription
record_clip_and_summarize	Screen clip with keyframe extraction
summarize_screenshot	Analyze an existing image file
transcribe_audio	Local Whisper transcription
get_recent_context	Retrieve recent Context Pack summaries

Key features

• Multi-monitor: active, specific, or all monitors at once
• Multi-device: Android screen via ADB
• Remote: SSE transport with bearer token auth
• Voice + visual capture simultaneously
• Annotation overlay (arrows, circles, rectangles, freehand, text)
• Vision providers: Ollama (free/local), Gemini (free tier), OpenRouter, Hyperbolic
• pip install eye2byte

Homepage: https://github.com/wolverin0/Eye2byte

[openclaw-barnacle]

Thanks for the pull request! This repository is read-only and is automatically synced from https://clawhub.ai, so we can’t accept changes here. Please make updates on the website instead.