ClawHub
Back to skill

Security audit

StockStar LLM Ranking - 证券之星大模型调用排行榜

Security checks across malware telemetry and agentic risk

Overview

This skill fetches a public LLM ranking page and formats the results; its main issue is broad activation wording, not hidden or destructive behavior.

Install this if you want an agent to answer Chinese-language questions about StockStar/OpenRouter LLM call rankings. Be aware it may contact tech.stockstar.com when activated, and the activation phrases should ideally be narrowed so ordinary GPT/Claude/model discussions do not trigger an unnecessary lookup.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • 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
Findings (3)

Lp3

Medium
Category
MCP Least Privilege
Confidence
90% confidence
Finding
The skill explicitly instructs the agent to fetch live data from an external website via a Python CLI, which implies network access, but no permissions are declared. Undeclared network capability weakens policy enforcement and user/admin visibility into what the skill can access, creating a transparency and control gap rather than a direct exploit by itself.

Vague Triggers

Medium
Confidence
88% confidence
Finding
The trigger list includes broad generic terms like '大模型', 'Tokens', '调用量', and vendor names such as 'GPT' and 'Claude', which can cause the skill to activate for unrelated conversations. Over-broad activation can route user requests to an external-data-fetching skill unexpectedly, causing unnecessary network calls, incorrect tool selection, and possible leakage of user intent to third-party sites.

Vague Triggers

Medium
Confidence
91% confidence
Finding
The activation section states when the skill should trigger but does not define boundaries for non-applicable cases, increasing the chance of accidental invocation. In this skill, accidental invocation is more significant because activation leads to external website access and summarization of live data, so misrouting can expose user queries externally and degrade agent reliability.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.