We actively support the latest stable release:

⚠️ WARNING: This bot performs real-time financial trading using AI algorithms and live market data. Trading in financial markets involves significant risk, including the potential loss of capital.

By using DART, you agree to the following:

• Use the bot at your own risk.
• You are solely responsible for any trades executed by the bot.
• You must comply with all applicable regulations, exchange terms of service, and legal requirements in your jurisdiction.
• The maintainers of this project provide no warranty or guarantee of profitability, uptime, or data accuracy.
• This software is experimental and intended for educational and research purposes only.

Because DART operates in live environments with AI, WebSocket streams, and trading APIs, it is vulnerable to the following:

• API key or session token leakage
• Trade hijacking or spoofing
• Adversarial AI inputs or model poisoning
• Man-in-the-middle attacks over WebSockets
• Insecure GUI or code injection
• Risky model self-modification (code updates via LLMs)

• Store secrets in .env or use a secret vault.
• Do not hard-code any sensitive keys.
• Never commit authentication tokens to Git.

• Always use wss:// and verify SSL certificates.
• Validate and sanitize all inbound and outbound JSON.
• Limit API usage scopes and rotate keys frequently.

• Separate testing and production data environments.
• Sanitize all streamed or stored inputs into InfluxDB/VectorDB.
• Monitor memory usage, as live model learning can be data-intensive.

• Disable auto-retraining in production unless signed and verified.
• Avoid injecting untrusted data into LLMs or reinforcement learners.
• Implement logging and rollback for model updates.

• Lock local dashboards to localhost if not intended for web access.
• Implement rate-limiting and access control if deployed on a network.
• Disable code execution from the frontend UI.

We value the security community. To report a vulnerability:

1. DO NOT include any sensitive information when creating a public GitHub issue.
2. Include:
   • Description and severity
   • Reproduction steps
   • Suggested mitigation (if any)

We aim to respond in 72 hours and patch critical issues within 7 business days.

• ✅ bandit – static security scanner for Python
• ✅ safety – dependency vulnerability checker
• ✅ dotenv – for secure config management
• ✅ pyright – type checks to catch unsafe logic
• ✅ pytest – test framework for code integrity

None at this time in the current release.
Check the issues page for updates.

Use this software wisely, securely, and ethically.
By running DART, you accept full responsibility for your trading outcomes and usage behavior.