Theta Trading System

• Outcome: Review
• Latest audit: May 29, 2026 at 9:27 PM UTC
• Version:
  1.2.0

Overview

This looks like a stock-analysis tool rather than malware, but its financial performance claims are overstated and inconsistent with the included code.

Treat this as experimental research code only. Do not rely on the advertised 100% accuracy, weekly return, or automatic-evolution claims for real trades; run it in an isolated environment, inspect or change the hard-coded paths, verify dependencies and data freshness, and independently validate any outputs before acting on them.

SkillSpector

By NVIDIA

Vulnerability Patterns

• MCP Tool Poisoning: Hidden Instructions, Unicode Deception, Parameter Description Injection
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access
• Supply Chain: Unpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (4)

Tp4

High

• Category: MCP Tool Poisoning
• Confidence: 95% confidence
• Finding: The skill makes strong claims such as '100% accuracy,' hourly auto-evolution, multi-source fallback, and Ridge-based behavior, while the analyzed behavior reportedly uses different models, lacks the stated fallback and validation workflows, and includes unrelated packaging/release actions. This mismatch can mislead users into overtrusting automated financial recommendations and running code under false assumptions about reliability, scope, and operational safeguards.

Intent-Code Divergence

Medium

• Confidence: 91% confidence
• Finding: The README presents materially inconsistent statements about the system's current model and feature set, claiming v1.2.0 uses a Ridge model with 100% accuracy while another core section describes GradientBoosting with 14 features. In a trading skill, this can mislead users about system behavior, validation status, and risk characteristics, causing unsafe reliance on inaccurate documentation for financial decisions.

Description-Behavior Mismatch

Medium

• Confidence: 97% confidence
• Finding: The file implements only a simple heuristic ranker over local SQLite data, while the skill metadata markets it as a '100% accurate', auto-evolving, real-time, multi-source trading system. In a financial decision-support context, this mismatch is dangerous because users may place undue trust in outputs that are not backed by the claimed sophistication, validation, or live-data safeguards.

Intent-Code Divergence

Low

• Confidence: 85% confidence
• Finding: The script prints '股票推荐' (stock recommendations) even though it later warns that the system is a simplified reference scorer. In the context of a trading skill, presenting heuristic output as a recommendation can mislead users about the authority and reliability of the result, increasing the chance of unsafe financial decisions.

VirusTotal

67/67 vendors flagged this skill as clean.