# The 5-Minute Morning Briefing Setup Checklist

**A step-by-step architecture checklist for OpenClaw agents that deliver a morning briefing before you wake up.**

---

## What You'll Need

### 1. Core Files (5 files)

- [ ] **SKILL.md** — Your briefing skill definition (prompt, data sources, output format)
- [ ] **briefing-data.py** — Python script that fetches weather, markets, news
- [ ] **calendar integration** — Google Calendar API or equivalent for today's schedule
- [ ] **cron configuration** — Scheduled job that fires at your wake time
- [ ] **messaging config** — Telegram/Discord channel for delivery

### 2. Data Sources (pick your stack)

- [ ] **Weather** — wttr.in (free, no API key) or Open-Meteo
- [ ] **Calendar** — Google Calendar API (OAuth2 setup required)
- [ ] **News** — Brave Search API or RSS feeds for your niche
- [ ] **Markets** — Yahoo Finance (yfinance, free) or Alpha Vantage
- [ ] **Tasks** — Todoist API, Apple Reminders, or plain text file

### 3. Infrastructure

- [ ] **OpenClaw** installed and configured (v1.0+)
- [ ] **Python 3.10+** with requests/yfinance packages
- [ ] **Cron daemon** running (systemd, launchd, or OpenClaw cron)
- [ ] **Telegram bot token** (or Discord webhook) for message delivery
- [ ] **API keys** for chosen data sources (most are free tier)

---

## Setup Steps

### Step 1: Define Your Briefing Sections
Write down the 4-6 sections you want every morning. Common choices:
- Today's calendar (meetings, hard stops)
- Weather forecast
- Top 3-5 news stories in your niche
- Market snapshot (if relevant)
- Priority tasks
- Proactive recommendations

### Step 2: Create the Skill File
Define a SKILL.md that:
- Specifies the exact output format (sections, order, max length)
- Lists data source parameters
- Sets the delivery channel and schedule
- Includes customization instructions

### Step 3: Write the Data Fetcher
Create a Python script that:
- Calls each API (weather, calendar, news, markets)
- Returns structured data the skill can format
- Handles API failures gracefully (briefing still delivers with partial data)
- Caches responses to avoid rate limits

### Step 4: Wire Up the Cron Job
Schedule the briefing to fire 15 minutes before your wake time:
- OpenClaw cron: `{ "kind": "cron", "expr": "45 6 * * *", "tz": "America/Denver" }`
- Verify it fires by checking gateway logs the next morning

### Step 5: Test and Iterate
- Run the skill manually first: check output quality
- Let it run for 3 days before tweaking
- Adjust sections, add/remove data sources based on what you actually read
- Pro tip: shorter is better. If you skip sections, cut them.

---

## Common Pitfalls

| Pitfall | Fix |
|---------|-----|
| Briefing too long to read | Cut to 4 sections max. 200 words per section. |
| API rate limits hit | Cache responses. Use free tiers wisely. |
| Calendar auth breaks | Use service account, not user OAuth. Test weekly. |
| Briefing feels generic | Add personal context: "kids pickup at 3:30", "gym at 8" |
| You stop reading it | It's too long. Cut it in half. |

---

## Time Estimate

| Task | DIY Time | With Template |
|------|----------|---------------|
| Write skill prompt | 45 min | 5 min |
| Build data fetcher | 2-3 hours | 0 (included) |
| Set up calendar API | 1-2 hours | 10 min (guide included) |
| Configure cron + delivery | 30 min | 5 min (config included) |
| Test and debug | 1-2 hours | 15 min (tested on Pi) |
| **Total** | **5-8 hours** | **35 minutes** |

---

## Want the Full Implementation?

This checklist shows you the architecture. The **Morning Briefing Blueprint** gives you the complete implementation:

- ✅ Battle-tested SKILL.md (runs daily on a Raspberry Pi)
- ✅ Python data fetcher (weather, markets, news, calendar)
- ✅ Google Calendar API integration with setup guide
- ✅ Cron configuration (copy-paste ready)
- ✅ Step-by-step setup instructions
- ✅ Tested on real hardware, not a tutorial

**[$9 — Morning Briefing Blueprint →](https://aaronmakelky.com/shop.html)**

Runs on OpenClaw. Works on a Pi. Ships to Telegram. 35-minute setup.

---

*Built by [Aaron Makelky](https://aaronmakelky.com) — solo operator running AI agents on a Raspberry Pi since 2025.*