Morning Briefing

Setup

1. Install this skill to Skills/morning-briefing/
2. Create a scheduled agent to run it at your preferred morning time
3. Optional dependencies (the skill adapts to what's available):

• Data/tasks.db (DuckDB) with a tasks table for task tracking. If you use the task-manager skill, this exists already. If not, the tasks section is skipped.
• Gmail integration (connected via [Settings > Integrations](/?t=settings&s=integrations))
• Google Calendar integration
• Supermemory for meeting prep context

How to Run

This is an orchestration skill. Follow these steps in order, then compose and deliver the briefing.

Step 1: Gather Today's Tasks

If Data/tasks.db exists, run the helper script:

python3 /home/workspace/Skills/morning-briefing/scripts/tasks_query.py

Returns JSON with overdue and due_today arrays, sorted by priority.

If the database doesn't exist, skip this section.

Step 2: Gather Today's Calendar

Use the Google Calendar tool:

use_app_google_calendar(
  tool_name="google_calendar-list-events",
  configured_props={
    "calendarId": "primary",
    "timeMin": "T00:00:00",
    "timeMax": "T23:59:59",
    "timeZone": "",
    "singleEvents": true,
    "orderBy": "startTime"
  }
)

Replace ` with current date, with the user's UTC offset, and ` with their timezone.

Step 3: Gather Inbox Highlights

Use Gmail to find important unread emails:

use_app_gmail(
  tool_name="gmail-find-email",
  configured_props={
    "q": "is:unread is:important newer_than:1d",
    "withTextPayload": true,
    "maxResults": 5
  }
)

If fewer than 3 results, also try:

use_app_gmail(
  tool_name="gmail-find-email",
  configured_props={
    "q": "is:unread newer_than:1d -category:promotions -category:social",
    "withTextPayload": true,
    "maxResults": 5
  }
)

Step 4: Gather News Digest

Customize these searches for the user's interests. Examples:

Tech/AI news:

web_search(
  query="AI technology news highlights",
  time_range="day",
  topic="news"
)

Industry-specific news (customize for user's industry):

web_search(
  query=" news",
  time_range="day",
  topic="news"
)

Step 5: Compose the Briefing

Write the briefing in this format:

# Morning Briefing -- , 

## Today's Schedule

## Tasks
### Overdue

### Due Today

## Inbox Highlights

## News Digest
### 

### 

Add a one-liner observation at the top if something stands out (overdue count, big meeting, breaking news).

Step 6: Deliver

Email:

send_email_to_user(
  subject="Morning Briefing -- ",
  markdown_body=""
)

SMS (condensed version, max 3-4 lines):

send_sms_to_user(
  message="Morning:  tasks due,  overdue. . "
)

Step 7: Archive

Save the full briefing to:

/home/workspace/Records/Briefings/-morning.md

Customization

• News categories: Edit step 4 to match the user's interests and industry
• Delivery timing: Adjust the scheduled agent time
• Delivery channels: Email only, SMS only, or both
• Meeting prep: If Supermemory is available, search for context about upcoming meeting attendees
• Tone: Match the user's AI persona voice

Notes

• If any section fails to load (API error, no results), include it with a "Could not load" note rather than skipping entirely.
• News items should be deduplicated across sources.
• The SMS should be ruthlessly brief. Only the most critical info.

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: zocomputer
• Source: zocomputer/skills
• License: MIT

Install and usage instructions live in the source repository linked above.