avifenesh

Atrium — AI skill for Claude Code

AI community

Zero-dependency, single-machine life dashboard — GitHub tasks, local AI agents, system health, email/calendar, subscriptions — with a global quiet switch, a pluggable collector model, and an MCP serve.

How to install Atrium

This entry records only its repository, not the path inside it, so there is no exact command to give. Open avifenesh/atrium and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

What Atrium does

Zero-dependency, single-machine life dashboard — GitHub tasks, local AI agents, system health, email/calendar, subscriptions — with a global quiet switch, a pluggable collector model, and an MCP server. Loopback-only by design.

Alternatives in AI

  • J Space Cognition Suite V3.7 — J-Space Cognition Suite V3.7 - AI cognitive-enhancement Skills based on Anthropic's J-space global workspace r 3k ★
  • Complete Context For Vertex AI: Life Sciences MCP Plugin Generation — COMPLETE Context for Vertex AI: Life Sciences MCP Plugin Generation skill 1.7k ★
  • Investment News — 为 A股投资者打造的全球产业链资讯看板 · 12 大赛道一一对应 A股板块(半导体/AI/机器人/新能源车…),覆盖 100+ 权威源,用你自己的大模型每日提炼中文「今日要点」+ 翻译 · 全程本地、零 API key 447 ★

README

atrium

A zero-dependency, self-hosted life dashboard for one machine and one human: GitHub tasks, local AI agents, system health, email/calendar, subscriptions, cron, and backups on one screen, with an MCP server so an LLM can read it too.

A small Node daemon polls the things already running on your box into a single in-memory snapshot, serves it over REST + SSE to a web UI, and exposes it to LLMs over MCP. It is **loopback-only by design**: no auth layer, no cloud, no account: it binds `127.0.0.1`, refuses to start on any non-loopback host, and your data never leaves the machine. One idea throughout: anything noisy can be muted, visually (`ui`) or for real (`enforced`: the source itself is paused).

![Atrium Today workspace with example data](docs/img/now-view.png)

*Today, shown with example data. Saved threads, recent projects, attention, and the day ahead.*

**Stack:** Node ≥ 24 · TypeScript · React · systemd user service · MIT licensed · zero runtime dependencies.

Why this project

  • Use it when you live across GitHub, local AI agents, and a handful of services, and want one calm screen instead of ten tabs: without sending anything to a cloud dashboard.
  • Use it when you want a dashboard you fully own: it runs on loopback as a systemd user service, reads only local files and CLIs, and keeps secrets out of every surface.
  • Use it when you want an LLM (Claude Code, Cursor, any MCP host) to answer "what needs my attention?": atrium exposes the whole snapshot over MCP with read-only tools it can auto-run.
  • Use it when you want to bolt your own data source onto a dashboard: a plugin collector is one file and renders with no UI code.

Installation

Requires **Node ≥ 24** and a **Linux user systemd session**.

git clone https://github.com/avifenesh/atrium && cd atrium
./scripts/install.sh        # npm install + build + generate & start the systemd user unit
# open http://127.0.0.1:5599

`install.sh` renders `scripts/atrium.service