用 Codex 或 Claude 帮你安装 复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它检查 Skill 页面并帮你完成安装。
直接命令不会经过审查 Prompt;运行前请先检查来源。
npx skills add https://github.com/SharpAI/DeepCamera --skill homeassistant-bridge命令会保持在同一行。复制前请横向滚动并检查完整内容。
想先保存到本地?可下载 SkillsMP 当前能够提供的文件。
正在显示 SKILL.md
基于 SOC 职业分类
| name | homeassistant-bridge |
| description | Bidirectional Home Assistant integration — HA cameras in, detection results out |
| version | 1.0.0 |
| parameters | [{"name":"ha_url","label":"Home Assistant URL","type":"url","default":"http://homeassistant.local:8123","required":true,"group":"Connection"},{"name":"ha_token","label":"Long-Lived Access Token","type":"password","required":true,"group":"Connection"},{"name":"cameras","label":"HA Camera Entities","type":"string","placeholder":"camera.front_door, camera.backyard","description":"Comma-separated HA camera entity IDs","group":"Cameras"},{"name":"poll_interval","label":"Frame Capture Interval (seconds)","type":"number","min":1,"max":60,"default":5,"group":"Processing"}] |
| capabilities | {"ha_bridge":{"script":"scripts/bridge.py","description":"Bidirectional HA ↔ Aegis integration"}} |
Bidirectional integration: HA camera feeds flow into Aegis's AI pipeline (detection → VLM → REID), and analysis results flow back as HA image_processing entities for automations.
Home Assistant SharpAI Aegis
────────────── ──────────────
camera.front_door ───────────► Detection → VLM → REID
camera.backyard ↓
Analysis results
image_processing.aegis_* ◄────── → objects, descriptions
→ HA automations → person identity
→ notifications → smart alerts
This skill builds on DeepCamera's src/home-assistant-py/ and src/home-assistant-nodejs/ modules.
python3 -m venv .venv && source .venv/bin/activate
pip install -r requirements.txt