ClawHub

Wangwei Mkdir

Security checks across malware telemetry and agentic risk

Overview

The skill is poorly and confusingly documented, but the included scripts only print static text and show no file changes, credential access, network use, or persistence.

Install only if you are comfortable with a low-quality skill whose documentation does not match its behavior. Do not rely on it for mkdir, cron, or backup functionality; based on the reviewed files, it appears to only print repeated text.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • 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)

Description-Behavior Mismatch

Medium
Confidence
94% confidence
Finding
The manifest says the skill echoes user input, while the body describes backup/mkdir-style operations. This semantic mismatch can lead an agent or user to authorize execution under false assumptions, potentially triggering filesystem modifications when only harmless text output was expected.

Intent-Code Divergence

Medium
Confidence
88% confidence
Finding
The documentation labels actions as 'echo' but the comments describe directory backup behavior and invoke different scripts. This inconsistency is dangerous because it obscures what the skill actually does, making review and safe use harder and increasing the risk of accidental execution of unintended operations.

Vague Triggers

Medium
Confidence
82% confidence
Finding
The manifest description and introductory text are too vague to define the skill’s intended use or operational boundaries. Ambiguity in a skill that may touch the filesystem makes it easier for agents to select it inappropriately and harder for reviewers to assess whether its actions are expected and safe.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal