Convert camera input into concise environmental scene descriptions.
npx skills add https://github.com/alexaundre/mycc --skill cc-eye
Please help me install this Agent Skill. Skill: cc-eye Source: https://github.com/alexaundre/mycc/tree/main/.claude/skills/cc-eye Command: npx skills add https://github.com/alexaundre/mycc --skill cc-eye
cc-eye helps Claude Code quickly interpret current physical surroundings by processing camera data and producing concise scene descriptions, enabling situational awareness without manual input.
启动 cc-eye 守护进程后，使用触发词获取当前场景描述。
High-intent search queries and answers about installing and using this skill.
You can achieve real-time scene awareness by running a local perception daemon that processes camera input into concise environmental descriptions. This setup allows your Claude Code agent to monitor nearby objects, people, and activities without manual input.
To set up local camera scene description processing, you need a local camera setup, a moondream-capable model backend, and a running daemon. This configuration ensures safe local processing to preserve privacy and minimize latency.
A vision daemon handles live scene descriptions by continuously processing camera data in the background and caching scene snapshots. This allows AI agents to fetch the latest scene details quickly using natural language triggers.
Yes, you can use natural language triggers like /look or 看一眼 to fetch current camera scene descriptions. The daemon processes these triggers to return the latest cached environmental details to your agent.
Local multimodal scene processing does not require an internet connection because all camera data and model inference happen locally. This approach preserves privacy and minimizes latency for environmental awareness tasks.
markwang2658Deploy LLaVA for image-based conversations and visual question answering.
kandengConvert natural language commands into Aiya humanoid robot movement, gesture, and speech actions.
curiositechRun background AI agents that observe editor activity and offer contextual assistance.
EveryIncGenerate and edit images via the Gemini API with multi-turn refinement.
