Shadow Binance Bot

• Outcome: Pass
• Latest audit: May 29, 2026 at 8:35 PM UTC
• Version:
  1.3.3

Overview

This is a read-only Binance analysis skill that uses sensitive Binance credentials but does not place trades, withdraw funds, persist in the background, or send data to unexpected services.

Install only if you are comfortable letting the skill read Binance balances and trade history. Use a read-only Binance API key, disable trading and withdrawals, restrict the key by IP where possible, prefer environment variables over config.env, and treat all strategy simulations as educational estimates rather than trading advice.

SkillSpector

By NVIDIA

Vulnerability Patterns

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

Findings (5)

Intent-Code Divergence

Medium

• Confidence: 98% confidence
• Finding: The simulated PnL formula adds losses as positive contributions instead of subtracting them, which can materially overstate profitability and mislead users into believing a strategy is safer or more profitable than it is. In a trading-analysis skill, this is dangerous because users may act on distorted performance projections and take inappropriate financial risk.

Intent-Code Divergence

Medium

• Confidence: 98% confidence
• Finding: This reduced-trading simulation repeats the same arithmetic flaw by adding loss amounts rather than subtracting them, producing inflated or inverted results. Because the function presents the output as a realistic improvement estimate, users could be nudged toward risky trading decisions based on false analytics.

Intent-Code Divergence

Medium

• Confidence: 97% confidence
• Finding: The DCA simulation claims reduced losses, but it multiplies the loss value and adds it to PnL, which can incorrectly increase reported returns if avgLoss is positive-coded. This creates misleading performance claims for a speculative strategy and may encourage users to adopt riskier behavior under false assumptions.

Missing User Warnings

Medium

• Confidence: 90% confidence
• Finding: The README instructs users to connect a Binance account and retrieve trading history, and it explicitly integrates AI/LLM components, but it does not clearly disclose what account/trade data may be transmitted, stored, or exposed to third-party services. In a finance-related skill, this omission is risky because trade history, balances, and API credentials are sensitive, and users may not understand the privacy and security implications of enabling the integration.

Missing User Warnings

Medium

• Confidence: 83% confidence
• Finding: The code generates strategy recommendations and improvement claims for DCA and related trading approaches without clearly warning that the outputs are hypothetical, simplified, and not financial advice. In a trading context, this omission increases the chance that users will overtrust simulated results, especially when the same file already contains misleading PnL logic.

VirusTotal

64/64 vendors flagged this skill as clean.