Trading Hub
Security checks across static analysis, malware telemetry, and agentic risk
Overview
This package is a full-featured autonomous trading bot whose code and runtime instructions request and use local credentials and register recurring autonomous tasks, but it misstates what it requires and contains prompt-injection signals — proceed only after careful review and containment.
Plain-language next steps and cautions: - This skill is a full autonomous trading bot (scanners, auto-trader, binance API integration) that will register recurring cron jobs to run trades and post reports. Do NOT install it into a production agent without review. - Key mismatch: The skill declares no credentials required but its code reads ~/.openclaw/workspace/USER.md for API keys/secrets. Before installing, inspect USER.md and remove any high-privilege keys or ensure it does not contain secrets you care about. - If you plan to test: use strictly sandboxed credentials (testnet keys or a dedicated sub-account) with minimal permissions (disable withdrawals, set tight trade limits, low balance). Never provide your main exchange API keys. - Disable or do not run the openclaw cron add commands until you audit the scripts. The SKILL.md includes explicit commands that will create scheduled autonomous tasks; run these only after code review and while in an isolated test agent. - Audit the code paths that send data out: check who receives announcements (the cron messages include a hard-coded Feishu user identifier). Verify that reports and logs are not being sent to unknown third parties and remove or change hard-coded recipients. - Because the SKILL.md contains detected unicode-control-chars (possible prompt injection), open the SKILL.md source in a safe text editor and verify there are no hidden instructions or maliciously crafted control sequences. - Run the bot in mock mode first (mock_trade.py / mock_trade_live.py) and review logs in scripts/data before allowing any real trading. Enable a 'kill switch' (e.g., set auto_trade_enabled=false in auto_config.json and kill cron jobs) as a safety precaution. - If you lack code-audit skills, do not install or run this skill with real funds. Consider asking a trusted developer/security reviewer to audit the code for credential exfiltration, unexpected network endpoints, or backdoors.
Static analysis
No static analysis findings were reported for this release.
VirusTotal
No VirusTotal findings
Risk analysis
No visible risk-analysis findings were reported for this release.