ClawHub
Back to skill

Security audit

Thai Chinese Gov Efficiency

Security checks across malware telemetry and agentic risk

Overview

This is mostly an academic research skill, but it tells agents to run local commands, spawn subagents, and update another skill on a recurring schedule without clear user approval or version pinning.

Install only if you are comfortable reviewing and approving any local scripts, R/Python commands, subagent delegation, and skill updates before they run. Do not allow the weekly update instruction unless you intentionally approve a pinned, reviewed version of the referenced skill.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • 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 (2)

Context-Inappropriate Capability

Medium
Confidence
88% confidence
Finding
The skill embeds executable commands and sub-agent spawning instructions, including direct use of `exec` and `spawn_subagent`, which expands its operational scope from academic analysis into code execution and orchestration. In a prompt- or tool-driven agent environment, this increases the risk of unintended command execution, misuse of local scripts, or lateral capability expansion beyond what is necessary for comparative education research.

Context-Inappropriate Capability

Medium
Confidence
84% confidence
Finding
The instruction to run `clawhub update --skill academic-research` introduces unnecessary external tool and package update behavior into a research skill. This can create supply-chain and environment-drift risk by encouraging retrieval and installation of unreviewed updates unrelated to the immediate academic task.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.