AI Agent Hub
Back to plugins
🎮

dsh-plugin-garmin

Fun & Dress-up Updated 2026.09.04

Run the following command in DeepSeek Harness:

dsh plugin install RayzPub/dsh-plugin-garmin

Paste the following prompt into your AI chat to install this plugin:

Clone the repository from https://github.com/RayzPub/dsh-plugin-garmin, build the source code, then run dsh plugin install RayzPub/dsh-plugin-garmin to mount the plugin into your DeepSeek Harness Profile.

About this plugin

Designing a custom watch face for the Garmin Fenix 7 used to mean hand-writing Monkey C, squeezing every byte of the 128 KB memory budget, manually aligning pointer pixels on the 260×260 round display, and repeatedly trial-and-erroring within the 64-shade MIP palette—one misstep and you get a face that simply will not render. dsh-plugin-garmin compresses that entire pipeline into a single sentence: describe the style you want in the DeepSeek Harness chat, and the plugin drives design, simulation, code generation, and firmware build in one pass.

Five registered tools power the plugin. garmin_specs surfaces the hardware constraints, garmin_preview renders a declarative WatchFaceSpec into a 260×260 SVG with memory estimation, garmin_scaffold generates a production-ready Connect IQ project with dynamic pointer trigonometry, safe sensor queries, and Sleep-Mode handling injected by the engine rather than hardcoded templates, garmin_build invokes the monkeyc compiler to emit a sideloadable .prg binary, and garmin_env checks or auto-initializes the Linux toolchain and 4096-bit RSA signing key. A built-in Euclidean-distance color-snapping algorithm guarantees every output shade stays within the 64-shade MIP legal range, eliminating non-renderable color errors.

It suits three audiences: end users who want a personalized Fenix 7 face without diving into the Connect IQ SDK; automation engineers who need to batch-generate watch-face firmware in CI or scripted pipelines; and AI Agent developers who want to drive hardware-level code generation through natural language. You describe the requirement; the plugin handles the rest.

Use Cases

  • Describe a tactical watch face in natural language and compile it into a .prg firmware
  • Batch-produce watch-face firmware in CI pipelines without hand-writing Monkey C
  • Preview 260x260 SVG in the Web Workbench, verify MIP color compliance, then export the project

Best For

  • End users who want a personalized Fenix 7 face without diving into the Connect IQ SDK
  • Automation engineers who need to batch-generate watch-face firmware in CI or scripts
  • AI Agent developers who want to drive hardware-level code generation and builds via natural language