ClawHub
Back to skill

Security audit

K8s容器编排工具

Security checks for vulnerabilities and agentic risk

Overview

This Kubernetes review skill is mostly coherent, but it asks for command and cluster access while using overly broad trigger language and only partial safety boundaries.

Install only if you intend to let the agent inspect Kubernetes manifests and, when you ask for it, run read-oriented kubectl commands against a configured cluster. Review commands before execution, avoid pointing it at production or full-cluster scope by default, and treat exported YAML as sensitive because it may contain cluster configuration or Secrets.

SkillSpector

Vulnerability Patterns
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • 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)

Vague Triggers

High
Confidence
95% confidence
Finding
The skill description uses broad, generic activation language such as productivity, automation, and workflow optimization rather than tightly scoping the skill to Kubernetes review tasks. This can cause the agent to invoke a skill with read/exec capabilities in contexts unrelated to Kubernetes, increasing the chance of unintended command execution or access to sensitive local/cluster data.

Missing User Warnings

Medium
Confidence
88% confidence
Finding
The skill advertises command execution and file handling capabilities but does not place prominent warnings, consent requirements, or safety boundaries around their use. In an agent setting, this can normalize impactful operations and make it easier for the skill to be invoked in ways that read local files, run shell commands, or touch cluster configuration without sufficiently informed user approval.

Static analysis

No suspicious patterns detected.