Back to skill

Security audit

Stock Expert

Security checks across malware telemetry and agentic risk

Overview

This stock-analysis skill is mostly purpose-aligned, but it publishes a real-looking Tushare API token and tells users to set it in their environment.

Review before installing. Do not use the published Tushare token; treat it as compromised and replace it with your own credential stored outside the skill files. The skill owner should revoke and rotate the exposed token, replace it with a placeholder, and add clearer credential-handling guidance. Only enable the cron examples if you intentionally want recurring stock reports.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (4)

Intent-Code Divergence

Medium
Confidence
91% confidence
Finding
The README states that the Tushare token is already configured while also publishing a concrete token value for users to set manually. This exposes what appears to be a live API credential and may mislead users into reusing or trusting embedded secrets, enabling unauthorized API use, quota theft, or account abuse.

Missing User Warnings

Medium
Confidence
83% confidence
Finding
The documentation instructs users to configure external market-data service tokens but does not clearly warn that these credentials grant access to third-party accounts and may be transmitted to external services during use. In an agent skill context, this increases the chance that users provide sensitive API keys without understanding the trust boundary or downstream data exposure.

Ssd 3

High
Confidence
99% confidence
Finding
A hardcoded Tushare API token is directly embedded in the setup instructions. This is a clear credential exposure issue: anyone reading the README can copy the token, consume the owner's API quota, impersonate usage, or potentially access account-associated data/services. The skill context makes this more dangerous because setup documentation encourages immediate reuse of the exposed secret.

Ssd 3

High
Confidence
99% confidence
Finding
The documentation includes a hardcoded Tushare API token in an environment-variable example, which can lead users or agents to reuse and disclose a real secret during installation, troubleshooting, or normal execution. In a skill that explicitly requires external market-data credentials, embedding a live token materially increases the chance of unauthorized API use, quota theft, account abuse, and secret propagation into logs, prompts, repositories, or model outputs.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.