ClawHub
Back to skill

Security audit

医械合规评分器

Security checks for vulnerabilities and agentic risk

Overview

This is a local medical-device compliance scoring tool, and the extra security-audit files do not add automatic network, credential, or privileged behavior.

Installers should treat this as a local self-assessment and report-generation aid, not a formal regulatory decision. Run the scripts only on scoring JSON you intend to process, and avoid opening generated HTML from untrusted third-party input because report fields are inserted into a local HTML file.

SkillSpector

Vulnerability Patterns
  • 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
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (1)

Tp4

High
Category
MCP Tool Poisoning
Confidence
94% confidence
Finding
The stated purpose is medical-device compliance scoring, but the finding indicates additional behavior that audits the skill package itself, scans for sensitive patterns, inspects external links, and emits a separate security_results.json. That is a material description-behavior mismatch: users may provide files expecting only product-compliance scoring, while the skill performs broader repository inspection and data analysis outside the declared scope.

Static analysis

No suspicious patterns detected.