agent-trading-predictor
Warn
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: MEDIUMCOMMAND_EXECUTIONREMOTE_CODE_EXECUTION
Full Analysis
- COMMAND_EXECUTION (MEDIUM): The skill utilizes the
mcp__flow-nexus__sandbox_executetool to run dynamically generated Python scripts. - Evidence: The 'Real-Time Trading Sandbox' section demonstrates executing a custom trading engine script provided as a string.
- REMOTE_CODE_EXECUTION (MEDIUM): Script execution in a sandbox environment constitutes remote code execution capabilities through the
mcp__flow-nexus__sandbox_executetool. - Evidence: Use of
mcp__flow-nexus__sandbox_executeto run logic containing infinite loops and simulated network connection calls (connect_market_feeds).
Audit Metadata