ClawHub
Back to skill

Security audit

Interface Health

Security checks for vulnerabilities and agentic risk

Overview

This is a useful network diagnostic guide, but it is labeled read-only while including commands that can reset counters or disrupt live interfaces.

Install only if users understand this is not purely read-only in practice. Use read-only accounts for diagnosis, preserve counter snapshots, and require human change-control approval before clearing counters, shutting interfaces, or re-enabling ports.

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 (3)

Description-Behavior Mismatch

Medium
Confidence
95% confidence
Finding
The skill metadata and description present the capability as read-only, but the troubleshooting section recommends `shutdown` / `no shutdown`, which is a state-changing operational action. That mismatch can cause an agent or operator to execute disruptive remediation under the assumption the skill is safe for observation-only use, creating avoidable service interruption risk.

Missing User Warnings

Medium
Confidence
97% confidence
Finding
The documented recovery step instructs operators to bounce an interface with `shutdown` / `no shutdown` without an explicit warning that this will interrupt traffic and may worsen incidents on production links. In a network operations context, such a disruptive action embedded in troubleshooting guidance is dangerous because it may be copied verbatim during urgent diagnosis without change-control review.

Missing User Warnings

Low
Confidence
80% confidence
Finding
The reference documents destructive counter-clearing commands without an immediate warning that they alter operational visibility by resetting diagnostic counters. In an agent-assisted workflow, an operator or downstream automation could treat all listed commands as equally safe to run, causing loss of forensic and troubleshooting context during incident analysis.

Static analysis

No suspicious patterns detected.