A Week with Muse as Daily Assistant: From Morning Briefings to Co-Creation

Connect Muse (Meta AI agent) to Google Calendar, GitHub, and other services; set up scheduled tasks and permission boundaries to automate daily summaries, information monitoring, content co-creation, and code review.

Who

Individual users already on Google Calendar who want to reduce repetitive checks and content production time

You get

Five automation scenes set up within a week; six note articles published in three days

You need

Muse · Google Calendar · GitHub · Google Sheets

Steps

1. Initial setup and service connections

   After registering Muse, complete base settings in Web or mobile app: name the AI (author chose 'Mocchi') and pick a persona tone (author set a casual, slightly cheeky character). Connect external services needed for automation: Google Calendar (for morning briefings), browser/search (for lottery monitoring), and GitHub (for code collaboration). In Settings → Data Controls, turn off the AI learning toggle to prevent conversations from being used for model training. On Android: open chat side panel → gear icon. On Web: Settings → Data Controls.

2. Configure and run daily automation tasks

   Describe each automation task to Muse and set triggers and filters: 1. Morning briefing: Schedule daily at 6:30. Muse reads Google Calendar, lists the day's events in chronological order, and adds one line of weather when available. Author reports Muse runs automatically at the set time and delivers the summary. 2. Pokémon card lottery monitoring: Schedule a midday check targeting two sites (Nyuka Now, cardchusen) with three exclusion rules: online or local-store channels only, exclude already-entered lotteries, exclude those requiring a purchase receipt. Muse notifies only items with approaching deadlines. 3. Co-writing note articles: Provide Muse with career history and episodes; Muse extracts potential topics, interviews to fill gaps, and generates draft copy plus header images. Author does final editing and presses publish. Six articles published in three days using this flow. 4. LINE sticker production: Maintain a Google Sheets table 'Stamp Neta Management' listing ideas, target sets, and status (not generated / generated / applied / published). When a set is ready, notify Muse, which invokes a self-made Python script to generate images. Author only reviews and submits to LINE Creators Market. 5. Development assistance: With GitHub connected, Muse can read all repositories. Author summarizes improvements as Issues, lets GitHub Copilot implement, and has Muse review—checking against instructions before human merge. Muse also clones source and runs it, removing the need for a cloud execution environment.

3. Set permission boundaries and safety habits

   Keep external-write, purchase, and public-publish actions on ask-each-time instead of always-allow. Principles: press important buttons yourself; correct factual errors on the spot; do not feed email or sensitive content for training; keep final judgment human. Lower-risk services (Rakuten, Amazon, etc.) may be set to always-allow to reduce friction. Periodically verify the AI learning toggle in Data Controls remains off.

No prompt needed — just follow the steps.