AI Library
Books for Reading AI
Choose a book, then read it in order from the table of contents.
[AI Library] Chapter 22: Morning Coffee Skill and Daily Task Automation
Mastering Claude Code
Chapter 22: Morning Coffee Skill and Daily Task Automation
Kim Kyung-jin
Mastering Claude Code
Introduction
8:47 a.m. While pouring a cup of coffee, I open VS Code and type a single line. "Make it morning today. Plan my day using the morning coffee skill." The agent starts moving. It reads the calendar, accesses the project management tool, and retrieves the quarterly goals file. By the time the coffee finishes brewing, the screen displays today's schedule arranged by the hour.
This is the morning coffee skill. The name is light, but what unfolds within it is quite intricate. It is an automation routine that weaves together three information sources,calendar, task management, and quarterly goals,into one, removing decision fatigue from the start of your day.
The Morning Briefing: Integrating Calendar, Project Management, and Quarterly Goals
When the morning coffee skill is called, the agent checks three locations in sequence.
First, the calendar. It retrieves today's scheduled meetings, calls, and appointments. If integrated with Google Calendar, it reads the schedule data through an MCP (Model Context Protocol) server or API. The goal of this stage is to determine what meetings are scheduled at what times and where the free time slots are.
Second, the project management tool. It logs into ClickUp or Notion to check this week's task list and status. It distinguishes between tasks with approaching deadlines, tasks in progress, and tasks not yet started. The agent does not judge this information on its own. It assigns importance based on the priority criteria written in context/priorities.md.
Third, quarterly goals. The context/goals.md file contains this quarter's key objectives and milestones. The agent cross-references how much each current task contributes to these goals.
[Figure 22-1] Integration diagram of three information sources in the morning coffee skill
The morning briefing is the synthesis of these three layers of information. The briefing includes the following:
There is one thing worth noting. When the agent constructs the briefing, it first reads CLAUDE.md, then sequentially references me.md, work.md, team.md, and priorities.md following the paths specified there. Every context file in the project folder flows into a single briefing. This is the moment when the 'home' and 'life' created in Chapter 20 actually come to life.
The briefing is not a simple list. It includes judgments like this: "We're at 60% of this quarter's goal achievement, but if we complete this task today, we can raise it to 70%." It is a mechanical application of the criteria written in the priority file, yet it is work that would normally require a person to open the calendar, check the project board, retrieve the goal document, and mentally combine the information while switching between three apps.
Priority-Based Automatic Schedule Arrangement
Taking a step beyond the briefing is the automatic schedule arrangement feature. When you see the agent's proposed daily plan and respond, "Good, let's go with that," the agent directly blocks out time slots in the calendar.
To understand why this feature is powerful, recall the concept of decision fatigue. Repeating questions like "What should I do for the next 15 minutes?" and "What about an hour from now?" in the morning consumes energy by itself. The morning coffee skill handles all these decisions at once.
The logic of automatic arrangement works as follows.
1. Check confirmed meetings and appointments in the calendar. 2. Extract free time slots. 3. Assign tasks to time slots according to the priorities listed in priorities.md. 4. Allocate generous blocks for tasks requiring focused work and smaller blocks for brief administrative tasks.
[Figure 22-2] Example screen of priority-based automatic schedule arrangement
When you respond with "Good," the agent calls the Google Calendar API to register each block as an event. You can also decline or request modifications. If you say, "I want to do video work first in the morning," the agent adjusts the block arrangement.
In this process, the agent performs two things. One is synthesizing information, the other is executing. The work of synthesizing information and creating a plan, and the work of reflecting that plan in your actual calendar, happen within a single conversation. The gap between "thought" and "action" disappears.
Team Pulse Check
While the morning coffee organizes an individual's day, the pulse check reveals the status of the entire team at a glance.
Give the agent a single instruction: "Run a team pulse check. I want to see if we're on track this week and this quarter." The agent accesses the project management tool and collects progress on all initiatives. It retrieves the owner, status, and deadline of each task, then analyzes them in light of team member information from team.md and quarterly goals from priorities.md.
[Figure 22-3] Example pulse check results report
The results report contains information like this:
On busy days, important follow-ups easily slip through the cracks. You might miss an approval request a team member is waiting for while concentrating on video production. The pulse check skill catches these blind spots.
Glancing inside the pulse check skill, you can spot signs of optimization. When retrieving data from the project management tool, searching for the list ID each time consumes tokens. So we store the IDs of frequently queried lists directly in the skill file. This allows the agent to access those lists without searching, reducing both speed and cost.
It is also effective to delegate the search work itself to a helper agent. If the skill file specifies, "Pass this query to the ClickUp searcher agent and get the search results," you can prevent the main agent's context window from being flooded with search data. The helper agent filters the raw data and returns only the needed information.
The Power of Parallel Work: Running Four Agents Simultaneously
Let me broaden the perspective now. There are four tasks: morning coffee, pulse check, research, and visual asset creation. If executed one by one in sequence, they take time even if the agent is fast. Let's see what happens when we run all four simultaneously.
Open four agent tabs in VS Code and issue one instruction to each.
It takes less than 30 seconds to type those four lines. And about two minutes later, results come back in all four tabs.
[Figure 22-4] Screen layout showing four agents working in parallel
In the first tab, today's schedule is organized. It is a complete morning briefing that includes calendar events, urgent action items, and video pipeline status. When you respond "Yes," the schedule is blocked into the calendar.
In the second tab, a LinkedIn post is written in your voice, and a seven-slide carousel is saved in the projects/carousels/ folder. It even includes a CTA (Call To Action).
In the third tab, the status of all initiatives and items needing follow-up are listed.
In the fourth tab, a PNG image file is saved in the projects/visualizations/ folder. It is a visual representation with a person buried in work on the left and a person working calmly with an assistant on the right.
What if you handled the same work manually? Ten minutes to open the calendar, check the schedule, and cross-reference tasks. Twenty minutes to research and write. Fifteen minutes to review the project board and organize status. Twenty minutes to create visuals. Including context switching, it easily exceeds an hour.
Parallel execution is possible because each agent has its own independent context window. While one agent reads the calendar, another searches the web, and yet another calls the ClickUp API. They do not interfere with each other's work. At the same time, they all access the same project folder, so they share communication style from me.md and business information from work.md. It is a structure that secures both independence and consistency at once.
Once you experience this, it becomes difficult to return to the approach of "asking one agent to do everything." It is like running multiple tasks simultaneously on a single computer. Being able to multiply your workload without sacrificing quality,that is what we call leverage.
Closing
The Morning Coffee skill is useful on its own, but its true value emerges when woven together with other skills. For morning briefings to surface project status, you need integration with project management tools, and to reference quarterly goals, files in your context folder must stay current. Your assistant's precision ultimately depends on context quality.
How you structure files, organize information hierarchically, and connect with project management tools. These choices shape your context handling skill, and that skill determines your assistant's ceiling.
AI Expert Attorney Kim Kyung-jin
Specialist in AI Law and Policy · Former Member of the National Assembly · Author of Multiple Works
If this book has stayed with you, even briefly, please help the next story reach the world.
(Voluntary support account: Nonghyup 302-1096-0948-81 Account holder: Kim Kyung-jin)
Kim Kyung-jin
Attorney · Former Member of the National Assembly · AI Policy Researcher
© 2026 Kim Kyung-jin. All rights reserved.