daily-review

Daily Review Skill

Generate comprehensive daily performance reviews with AI coaching insights.

Features

Usage

# Run daily review for today
~/clawd/skills/daily-review/scripts/daily-review.sh

# Run for specific date
~/clawd/skills/daily-review/scripts/daily-review.sh 2026-01-15

Sample Output

🏆 Daily Performance Review - 2026-01-15
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

📬 COMMUNICATION
  • Emails sent: 6
  • Slack messages: 203
  • X.com mentions: 5

📅 MEETINGS (Fireflies - speaker verified)
  • CEO Chat (70 min)
  • Meeting with Perfectos (27 min)
  • US Squad Standup (27 min)
  Total: 3 meetings (~2.0 hrs)

💻 OUTPUT
  • Git commits: 6
  • Docs modified: 20
  • Messages to Ada: 73

⏱️ FOCUS TIME
  Screen Time: 9.7 hrs
  • Atlas: 203min
  • Slack: 163min
  • Telegram: 45min
  
  ActivityWatch: 8.5 hrs
  • Telegram: 120min
  • Ghostty: 90min
  • Chrome: 45min

Requirements

APIs & Services

• Gmail: Google Workspace service account or gog OAuth
• Slack: Slack API token (user_token for search)
• Fireflies: API key for meeting transcripts
• Google Drive: Service account for docs tracking

Tools

• Bird CLI: For X.com/Twitter (requires auth_token + ct0 cookies)
• ActivityWatch: Local app tracking (http://localhost:5600)

macOS (for Screen Time)

• SSH access to Mac
• get_screentime.py script for knowledgeC.db queries

Installation

1. Copy skill to your clawd workspace:

cp -r daily-review ~/clawd/skills/

2. Install dependencies:

# Bird CLI (on Mac)
cd ~/Code && git clone https://github.com/steipete/bird.git
cd bird && npm install && npm run build:dist

# ActivityWatch
# Download from https://activitywatch.net/

3. Configure secrets:

# Bird (X.com)
cat > ~/clawd/secrets/bird.env << 'EOF'
AUTH_TOKEN=your_auth_token
CT0=your_ct0
EOF

# Fireflies
echo "your_api_key" > ~/clawd/secrets/fireflies.key

# Slack
echo '{"user_token": "xoxp-xxx"}' > ~/clawd/secrets/slack-super-ada.json

4. Add cron job for daily 09:00 review:

clawdbot cron add --name "daily-review" --schedule "0 9 * * *"

Screen Time Query

The skill queries macOS Screen Time directly from knowledgeC.db:

SELECT 
  ZVALUESTRING as app,
  SUM(ZENDDATE - ZSTARTDATE) as seconds
FROM ZOBJECT 
WHERE ZSTREAMNAME = '/app/usage' 
AND date(ZSTARTDATE + 978307200, 'unixepoch') = '2026-01-15'
GROUP BY ZVALUESTRING
ORDER BY seconds DESC

Fireflies Speaker Verification

Meetings are verified by checking if user actually spoke (not just invited):

{
  transcripts(limit: 30) {
    title dateString duration
    sentences { speaker_name }
  }
}

Only meetings where speaker_name contains user's name are counted.

License

MIT