claude-tradingview-mcp-autotrading

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly instructs cloning and running code from the external GitHub repo during setup/one-shot execution (git clone https://github.com/jackson-video-resources/claude-tradingview-mcp-trading), which means remote code would be fetched and executed as a required dependency at runtime.

MEDIUM W009: Direct money access capability detected (payment gateways, crypto, banking).

• Direct money access detected (high risk: 1.00). The skill is explicitly a cryptocurrency auto-trading bot designed to place real market orders. It documents connecting to exchange APIs (BitGet, Binance, Bybit, OKX, Coinbase, Kraken, KuCoin, Gate.io, MEXC, Bitfinex), requires API credentials (API key, secret, passphrase) in .env, and includes code that calls exchange.createMarketOrder to execute live trades. It also describes live vs paper trading, VPS deployment for 24/7 execution, and safety rules for exchange API settings. These are specific, purpose-built financial execution capabilities (placing trades via exchange APIs), not generic tooling.

MEDIUM W013: Attempt to modify system services in skill instructions.

• Attempt to modify system services in skill instructions detected (high risk: 0.80). The skill explicitly instructs SSHing in as root, running apt installs, editing crontab, checking systemctl, and creating files under /root (cron jobs, service checks, VPS deployment), all of which modify system state and require elevated/sudo privileges on the host—so it pushes actions that can change the machine's state.