This project supports a personally-owned, 24/7 AI-model trading setup running on Windows 11. It covers:
- Google Workspace — docs, sheets, and email used for trade logs, reports, and research notes.
- Microsoft Edge — browser environment for broker dashboards, exchange consoles, and Workspace access.
- Windows 11 host — the machine running the trading model/server continuously.
- Home network (multiple SSIDs / WLAN, WLAN1, WLAN2) — separate wireless networks used to segment traffic (e.g., trading server on an isolated SSID, general devices on another).
- Local file agent — a command-line tool (
agent.py) for managing trading data on this machine: reading logs, creating new files/reports, compressing archives, and encrypting sensitive data at rest.
- Everything here operates on the local machine only. There is no remote-control, remote-monitoring, or cross-device command execution.
- Multi-device administration (if you manage more than one PC on your
network) should go through Windows-native tools:
- PowerShell Remoting (
Enter-PSSession,Invoke-Command) for scripted admin tasks on machines you own and have credentials for. - Group Policy or Microsoft Intune/Endpoint Manager for fleet-wide config and security policy.
- Windows Admin Center for a GUI-based multi-machine dashboard. These are Microsoft's supported, auditable paths for device administration and are safer/more maintainable than custom scripts.
- PowerShell Remoting (
- Network segmentation (separate SSIDs) is configured on your router/AP firmware directly — this repo doesn't need to touch that.
| Component | Role |
|---|---|
agent.py |
Local CLI: read, create, compress, encrypt files |
| Google Workspace | Reporting, shared logs, documentation |
| Edge | Browser access to broker/exchange platforms and Workspace |
| Trading server (Windows 11) | Runs the AI trading model continuously |
See agent.py. Commands:
python agent.py read <file>
python agent.py create <file> --content "..."
python agent.py compress <output.zip> <file1> <file2> ...
python agent.py encrypt <file> --password "..."
python agent.py decrypt <file.enc> --password "..."
- Encryption uses a password-derived key (PBKDF2 + Fernet/AES). Keep your password somewhere safe — there is no recovery if it's lost.
- Encrypted/compressed archives are written locally; nothing is transmitted over the network by this tool.
- Treat trading credentials and API keys separately (e.g., environment variables or a secrets manager), never hard-coded into scripts.