G2量化交易策略

Security checks across malware telemetry and agentic risk

Overview

This stock-trading skill is mostly coherent, but it asks for unrelated wallet/signing capability and contains under-scoped trading/data behaviors that users should review before installing.

Review before installing. Do not grant wallet or transaction-signing permissions unless the publisher clearly explains why a stock-analysis skill needs them and you can approve each action. Run only the included stock_analysis_v5.py in an isolated workspace first, treat the backtest claims as unverified financial marketing, and back up any OpenClaw finance database before using save or workflow features.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (1)

Intent-Code Divergence

Medium
Confidence
93% confidence
Finding
The configuration contains contradictory execution rules: elsewhere it states stop-loss/take-profit decisions should use closing prices only, but the monitoring section says some sell triggers should execute immediately or in real time. In an automated or semi-automated trading agent, this ambiguity can cause inconsistent behavior, unexpected premature liquidation, or bypass of intended risk controls depending on which rule the implementation follows.

VirusTotal

67/67 vendors flagged this skill as clean.

View on VirusTotal