ClawHub

Skill Quality Scorer

Security checks across malware telemetry and agentic risk

Overview

This is a local skill-quality scoring aid that reads target skill files and runs a bundled static audit script, with no evidence of hidden persistence, exfiltration, or destructive behavior.

Install only if you want an agent to inspect local skill directories for quality scoring. Run it on directories you intentionally choose, avoid pointing it at unrelated private workspaces, and treat its scoring as an advisory review aid rather than a substitute for behavioral testing or human security review.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • 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 (2)

Lp3

Medium
Category
MCP Least Privilege
Confidence
82% confidence
Finding
The skill instructs the agent to read files and invoke a shell script (`python scripts/static_audit.py`) but does not declare corresponding permissions. This creates a capability/permission mismatch that can lead to unintended file access or command execution if the platform relies on declared permissions for review, sandboxing, or user trust.

Tp4

High
Category
MCP Tool Poisoning
Confidence
92% confidence
Finding
The skill advertises a deterministic TRACE+ scorer with JSON and Markdown output, comparison workflows, and a full 30-sub-item rubric, but the analyzed behavior does not implement that claimed functionality. Security-wise, this is dangerous because users may overtrust audit results from a tool that only performs limited static checks, leading to false assurance and missed risks in downstream skill evaluation.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal