ClawHub
Back to skill

Security audit

A股模拟持仓分析

Security checks across malware telemetry and agentic risk

Overview

The skill is a coherent Eastmoney MX simulated-trading helper, with expected account API access and simulated-account mutations, but users should treat its API key and posting/trading actions carefully.

Install only if you intend to connect Codex to your Eastmoney MX simulated account. Keep MX_APIKEY private, leave MX_API_URL at the trusted default unless you know exactly why you are changing it, and review simulated buy/sell/cancel or post requests before asking the agent to run them.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Taint TrackingDirect Taint Flow, Variable-Mediated Taint Flow, Credential Exfiltration Chain
  • 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
Findings (3)

Tainted flow: 'cmd' from os.environ.get (line 26, credential/environment) → subprocess.run (code execution)

Medium
Category
Data Flow
Content
'-d', json.dumps(payload)
    ]
    try:
        result = subprocess.run(cmd, capture_output=True, text=True, timeout=30)
        if result.returncode != 0:
            print(f"[ERROR] curl failed: {result.stderr}")
            return None
Confidence
86% confidence
Finding
result = subprocess.run(cmd, capture_output=True, text=True, timeout=30)

Tp4

High
Category
MCP Tool Poisoning
Confidence
86% confidence
Finding
The documented scope says the skill should act only on explicit user requests for simulated trading operations, but the referenced behavior includes an automatic --auto-post workflow that can inspect account activity and initiate posting without a direct user request. That creates a trust-boundary violation: account-derived data may be processed or published unexpectedly, increasing the risk of unintended disclosure or unauthorized actions even in a simulated account.

Missing User Warnings

High
Confidence
97% confidence
Finding
Buy, sell, and cancel actions are executed immediately once the parser infers intent, with no explicit confirmation, preview, or secondary check. In an agent setting, natural-language misparsing, prompt injection, or accidental invocation could trigger destructive portfolio actions, making this especially dangerous even in a simulated trading context.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.