ClawHub
Back to skill

Security audit

Data Analyst Chinese Free

Security checks across malware telemetry and agentic risk

Overview

This is a disclosed Chinese data-analysis template skill with expected read, write, and execution needs, though users should be cautious with optional network features.

Install only if you want an agent to run Python-style data analysis in your workspace. Review any generated code before execution, keep sensitive datasets scoped to the intended files, and only use API URLs or callback_url values that you trust.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Behavioral ASTexec() Call, eval() Call, Dynamic Import
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Findings (4)

Description-Behavior Mismatch

Medium
Confidence
91% confidence
Finding
The skill is presented primarily as a local data-analysis template, but it also documents `requests.get(url).json()` for third-party API access. That expands the trust boundary and can cause outbound data transfer or retrieval from attacker-controlled endpoints without clear disclosure, which is a meaningful capability mismatch for a skill with read/write/exec tools.

Context-Inappropriate Capability

Medium
Confidence
95% confidence
Finding
Advertising a `callback_url` introduces outbound network interaction that is not necessary for a local analytics-template skill. If honored by an agent, it can be abused for data exfiltration, SSRF-like requests to internal services, or sending results to attacker-controlled infrastructure.

Vague Triggers

Medium
Confidence
84% confidence
Finding
The trigger condition ('use when data analysis, report generation, statistical insights, or visualization are needed') is broad enough to capture many generic user requests. In an environment with tool routing, overbroad triggers can cause this write/exec-capable skill to activate unnecessarily, increasing the chance of unintended file access, command execution, or mishandling of sensitive datasets.

Missing User Warnings

Medium
Confidence
93% confidence
Finding
The manifest enables `write` and `exec`, but the skill description frames itself as simple analysis templates without clearly warning that it may modify files or execute commands. This mismatch is dangerous because users may provide data expecting passive analysis while the agent has authority to alter the workspace or run code, which materially raises the impact of prompt injection or operator misunderstanding.

VirusTotal

62/62 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.