Sabrina banner
willyzanardo willyzanardo

Sabrina

AI community

Description

Local-first macOS assistant that keeps you focused on what you committed to finish today, while watching your AI coding agents and gathering Notion, Slack, Git, and calendar signals into your next priorities.

Installation

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

README

Sabrina 🦉

Sabrina is a local-first personal activity hub for macOS. She watches your AI coding agents (Claude Code and Codex), gathers everything that needs your attention — Notion tickets, Slack requests, Git state, upcoming meetings — and turns it into one focus board and one visible next action. A small animated owl floats above your other windows and tells you which agent is working, which one is waiting on you, and what you committed to finish today.

Everything runs on your Mac. Tokens live in the Keychain, raw prompts and messages are processed in memory and discarded, and only metadata, summaries, and references are persisted to disk.

![Sabrina's main window showing today's firm outcomes and the focus board (sample data)](docs/screenshots/main-window-focus.png)

The floating overlay: Sabrina's mascot, agent status capsules, and the focus summary (sample data)

*Both screenshots show sample data from fictional projects.*

What Sabrina does

  • Menu bar + window app for macOS 14+, in English and Spanish.
  • Firm daily outcomes: pick one to four outcomes across your projects, order them, and define what "done" means for the primary one. The board header holds you to that plan.
  • Agent awareness: per-session status for Claude Code and Codex — working, waiting for you, finished, session closed — with capsules in the floating overlay, local chimes, and native notifications when an agent asks for permission or finishes a turn. Clicking a capsule or notification focuses the exact Terminal or iTerm tab that owns the session.
  • Automatic agent context: sessions are matched to your daily outcomes from their prompt summary and working directory, so the overlay shows which outcome each agent is advancing. Ambiguity stays quiet; only a confident mismatch sustained for ten minutes is raised as possible divergence.
  • Deliberate completion check: a