Public agent skills for working with Honch — product analytics for connected hardware. These are drop-in skills for AI coding agents (Claude Code, Codex, Cursor, Gemini CLI, Copilot) that encode how to do a Honch job well, so the agent gets it right the first time.
| Skill | What it does |
|---|---|
honch-integrator |
Integrate Honch into any codebase — detect the right path (ESP-IDF, Arduino, C/POSIX, MicroPython, relay, or HTTP/JSON), wire it in, and verify a first event. Targets SDK 0.3.0. |
More skills (e.g. event instrumentation, integration diagnosis) may land here over time.
Each skill installs with one line, via an interactive selector that supports multiple agents (zero dependencies — no Node):
curl -fsSL https://honch.dev/skill | shSee honch-integrator/README.md for flags, scopes, and manual installation.
Skills are written agent-neutral and wrapped per agent:
<skill>/
content/ # the real instructions, in plain action-language (no agent-specific tools)
adapters/ # thin per-agent entry points (Claude SKILL.md, AGENTS.md, Cursor .mdc, …)
install/ # the installer the honch.dev route serves
VERSION
The knowledge lives once in content/; the adapters only handle discovery, so the
same skill works across agents. AGENTS.md is the cross-agent lingua franca.
A skill's VERSION tracks the Honch SDK version it targets. The installer fetches
a matching tag, so an install is pinned and ages honestly.