sensecraft-solutions

从原始资料创作一个符合 spec 的一键部署 IoT 方案。基于 Wiki/文档/Git 仓库复现方案，提炼最简路径，输出符合公开契约（spec/CONTRACT.md）的 solution.yaml / guide.md / description.md，并用离线工具 solutionctl 校验。适用于：从资料创建新方案、提炼最简部署路径、校验方案合规性。

Deploy an IoT solution via the provisioning station API. Use this skill whenever the user asks to deploy, flash, install, or set up any solution on devices — even if they don't say "deploy" explicitly. Also use it when troubleshooting a failed deployment or checking deployment status.

Prepare Debian packages for reCamera C++ deployment. Use when creating .deb packages for reCamera devices, configuring init scripts, or setting up binary deployment files.

Prepare ESP32 firmware files for solution deployment. Use when setting up firmware flashing, configuring esptool parameters, or adding ESP32/ESP32-S3/ESP32-C3 device support to a solution.

Prepare Himax WE2 firmware and AI models for SenseCAP Watcher deployment. Use when setting up xmodem flashing, configuring AI model addresses, or adding Himax device support to a solution.

优化 IoT 解决方案文案。检查并改进 solutions/ 目录下的介绍页和部署页文案，确保非技术用户能理解。使用场景：优化文案、检查术语、修复文案问题。

PREVIEW locally edited solution content inside the user's installed SenseCraft Solution desktop app, without rebuilding the app. This is **preview only — not a release**. Use when the user says "本地预览一下", "我改了文案，App 里能看到吗", "把这个方案导进 App 看看效果", or similar. The real release path is OTA (`scripts/generate_solution_manifest.py` + commit `bundled_hashes.json`) and is a separate, intentional action — if the user wants to ship to all users, do NOT use this skill.

教 AI agent 用 solutionctl CLI 调用引擎能力——发现方案、部署、校验、设备管理。当 agent 要在命令行 / CI / 无 GUI 环境部署或操作 SenseCraft 方案时加载。

Create a Jetson docker_remote solution package from structured inputs, with optional RTSP preview integration. Use when generating a one-click Jetson deployment solution from a Docker image and start command.

Prepare Docker images and compose files for solution deployment. Use when building container images, writing docker-compose.yml, configuring health checks, or setting up multi-container services.

Prepare Node-RED flows for reCamera deployment. Use when creating flow.json files, configuring InfluxDB connections, or setting up vision processing pipelines on reCamera.