QuantDinger

QuantDinger Security Hardening Guide

Least-privilege permissions

• Restrict API keys to essential actions only: Issue separate exchange API keys for read-only, paper trading, and live trading. Each key should grant only the minimum permissions needed (e.g., "place orders on BTC/USD pair only" rather than "all trading permissions").
• Limit agent access to specific strategies: Deploy agents with access only to the trading strategies they need to execute, not the entire codebase or admin functions.
• Rotate credentials regularly: Change API keys, database passwords, and service account tokens every 30 days; revoke any unused credentials immediately.
• Separate user roles and approval chains: Require manual approval from a human trader before live trades execute; prevent agents from accessing billing, user management, or settlement systems.

Network isolation

• Encrypt all data in transit: Use HTTPS/TLS for all connections between QuantDinger components, exchanges, and user devices; disable unencrypted protocols entirely.
• Isolate trading engines from the internet: Run core trading logic on private networks with controlled egress; only expose API endpoints through a secure gateway that filters requests and logs access.
• Segment live and paper trading environments: Keep test/backtest systems completely separate from production accounts to prevent accidental live trades during testing.
• Restrict third-party integrations: Only allow connections to approved exchanges and data providers; use API tokens with IP whitelisting and request signing when available.

Monitoring

• Log all trades and agent decisions: Record every order placed, strategy change, and approval decision with timestamps; retain logs for 12+ months for audit trails and post-incident investigation.
• Set up real-time alerts for anomalies: Alert traders immediately on unusual activity (e.g., trades 10x larger than typical, strategies accessing new markets, failed login attempts, credential usage from unexpected locations).
• Audit dependency updates: When updating Python libraries, scan new versions for known vulnerabilities using tools like OWASP or Snyk; test updates in paper trading before production deployment.
• Monitor resource usage and costs: Track API call rates, compute usage, and trading volume to detect compromised credentials or runaway agents draining accounts.

Red flags to watch for

• Unexpected trades or strategy behavior: If an agent places trades that don't match the configured strategy or exceeds position size limits, pause immediately and review audit logs for signs of compromise.
• Unusual credential activity: API keys being used from unexpected geographic locations, at odd hours, or with rapid-fire requests suggest theft; revoke them and rotate all secrets.
• Performance degradation or service outages: Sudden slowdowns or crashes may indicate a denial-of-service attack or malicious code injection; isolate affected systems and scan for unauthorized changes.
• Unaccounted data or financial discrepancies: Missing audit logs, unexplained fund transfers, or strategy files that have been modified without approval indicate a breach; trigger incident response and contact your exchange immediately.