ClawHub
Back to skill

Security audit

Skills of A-share market data released by ft.tech.

Security checks across malware telemetry and agentic risk

Overview

This skill appears to be a disclosed read-only market-data helper that runs bundled Python scripts to fetch public A-share stock data from stated domains.

Install only if you are comfortable with the agent running the bundled Python scripts and making requests to market.ft.tech and ftai.chat. Keep use to the documented subskills and parameters; the dispatcher does not enforce a separate whitelist, so operators in tightly governed environments may want additional command and domain allowlisting.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • 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)

Lp3

Medium
Category
MCP Least Privilege
Confidence
91% confidence
Finding
The skill instructs the agent to derive and execute a local `run.py`, optionally read sibling `sub-skills/.../SKILL.md` files, and call external HTTPS endpoints, which together imply shell execution, file reads, and network access despite no declared permissions. This creates a trust and containment gap: operators reviewing only declared permissions may not realize the skill can execute code and reach the network, increasing the chance of unintended command execution or data exfiltration if the implementation or sub-skills are compromised.

VirusTotal

59/59 vendors flagged this skill as clean.

View on VirusTotal