SEC 13F Whale Tracker

Track what Buffett, Dalio, Ackman, Soros, and Cathie Wood are buying and selling via SEC EDGAR 13F filings.

Setup (one-time)

SKILL_DIR="$(dirname "$(readlink -f "$0")")"  # or resolve from skill path
cd /tmp && python3 -m venv sec13f-venv
source sec13f-venv/bin/activate
pip install -r "$SKILL_DIR/scripts/requirements.txt"

Or reuse an existing venv if available. The only dependency is requests.

Run Tracker

python3 scripts/tracker.py

Output:

• data/ — cached holdings JSON per fund per quarter (auto-created)
• reports/latest.md — most recent report
• reports/13f-report-YYYY-MM-DD.md — dated report

Default Tracked Funds

To add funds, edit the FUNDS dict in scripts/tracker.py. Find CIK numbers at SEC EDGAR.

Report Structure

Each fund section includes:

• Report period and filing date
• Total holdings value and count
• 🆕 New positions with value and shares
• 📈 Increased positions with % change
• 📉 Decreased positions with % change
• 🚫 Closed positions

Cross-fund insights section highlights:

• Stocks where multiple managers converge (buy or sell simultaneously)

Cron Integration

Schedule quarterly runs (13F filings are due ~45 days after quarter end):

# Run mid-February, May, August, November
0 8 15 2,5,8,11 * cd /path/to/project && venv/bin/python scripts/tracker.py

Or use OpenClaw cron to run and send the report to a channel.

Technical Notes

• SEC EDGAR API is free and public; requires User-Agent header
• Rate limited to <10 req/sec (script adds 150ms delay between requests)
• Parses 13F information table XML, handles namespace detection
• Merges holdings by CUSIP, separates put/call options
• Caches data in data/ to avoid re-fetching known quarters
• No proxy needed for SEC EDGAR (script clears proxy env vars)