A Garmin AI Skill and Richer Health Tracking in Obsidian
Today I shipped a few things around health tracking in my Obsidian vault. They all click together.
All topics → news
news
Claude Code Routines
A Routine is a saved Claude Code configuration (prompt, repositories, environment, connectors) that runs automatically on Anthropic-managed cloud infrastructure. Routines generalize the earlier Claude Code Web Scheduled Tasks concept by adding multiple trigger types to the same saved session. Resear
Garmin AI Skill
An open-source AI Agent Skill (MIT License) that fetches health and exercise data from Garmin Connect and makes it available to AI agents. I built it to pipe data from my Garmin Venu 3 straight into my Obsidian daily notes.
Introducing DeveloPassion's News Flash
I've just launched a new News page on my blog and a matching newsletter: DeveloPassion's News Flash.
LiteLLM Claude Code Proxy
Open-source LiteLLM configuration that lets Claude Code run against an Azure AI Foundry deployment instead of Anthropic's hosted API. Published by National Bank of Belgium. It's a thin wrapper: a `litellmconfig.yaml` with a wildcard model route, a `startproxy.sh` launcher, a Dockerfile, and a `.env`
Sidenotes plugin for Obsidian
Adds Tufte-style sidenotes and margin notes rendered in the margin of an Obsidian note. Inspired by Gwern.net and Edward Tufte's CSS conventions. Combines a CSS snippet with a plugin to handle behavior CSS can't (editing in margin, re-sequencing, etc.).
LiteLLM
LiteLLM is an open-source AI Gateway that exposes 100+ LLM providers through a single OpenAI-compatible interface. It ships in two forms: a Python SDK for embedding the gateway logic inside applications, and a self-hostable proxy server (the "LLM Gateway") that centralizes access, spend tracking, an
My communications plan
This is my (Sébastien Dubois) Communications plan. It describes where/how I share information and ideas. It's part of my business flywheel.
Claude Code Prompt Caching
Claude Code uses Anthropic's prompt cache to reuse large, stable chunks of context (system prompt, tool definitions, CLAUDE.md, prior turns) across requests instead of resending and re-billing them every turn. Cache reads are cheaper than uncached tokens; cache writes are more expensive than normal
2026-04-14 Interesting Links
Here are a few links I found interesting today.