ClawHub
Back to skill

Security audit

Xskill Are

Security checks across malware telemetry and agentic risk

Overview

This is a documentation-only evaluator skill that uses disclosed rubrics and red-team test prompts to score other skills, with no installer code or hidden execution behavior found.

Before installing, understand that this skill is mainly Chinese-language and uses LLM judges, adversarial test prompts, and score storage as part of evaluation. It is suitable if you want a rubric-driven skill quality assessor, but treat scores as advisory because the artifacts themselves disclose limitations around language coverage, model drift, calibration, and judge subjectivity.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • 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
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Findings (1)

Natural-Language Policy Violations

Medium
Confidence
87% confidence
Finding
The rubric and judge prompt are written entirely in Chinese and appear to assume Chinese-language operation without offering a user-language choice or documenting a hard locale requirement. In an evaluation skill, this can degrade reliability, fairness, and auditability by systematically disadvantaging non-Chinese users or causing mis-scoring when prompts, capabilities, or outputs are in other languages.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

Detected: suspicious.prompt_injection_instructions

Prompt-injection style instruction pattern detected.

Warn
Code
suspicious.prompt_injection_instructions
Location
references/test-cases-en.md:26