ClawHub
Back to skill

Security audit

Docker Essentials V1 Tool Free

Security checks across malware telemetry and agentic risk

Overview

This Docker helper is mostly coherent, but it gives an agent broad command-line authority over containers, images, registries, volumes, and host package installation without enough scoping or confirmation guidance.

Install only if you want an agent to run Docker commands for a specific legacy Docker environment. Before using it, require explicit confirmation for any delete, prune, volume removal, image push, registry login, sudo command, or production service change, and provide exact container/image/project names rather than broad requests.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Behavioral ASTexec() Call, eval() Call, Dynamic Import
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Findings (6)

Description-Behavior Mismatch

Medium
Confidence
83% confidence
Finding
The skill markets itself as a basic Docker V1/container-management helper, but it also includes migration assessment, package installation, and broader environment-management steps. This scope expansion increases the chance an agent will perform higher-risk operational tasks than the user expects, including system modification and environment discovery.

Intent-Code Divergence

Medium
Confidence
86% confidence
Finding
The documentation says the skill is not appropriate for operations/deployment management, yet it provides docker-compose orchestration and service lifecycle commands that are operational in nature. This contradiction can mislead routing and safety policies, causing the skill to be invoked in contexts where destructive deployment actions are possible.

Vague Triggers

Medium
Confidence
79% confidence
Finding
The trigger condition is extremely broad, covering common coding, debugging, and deployment scenarios. In an agent environment, this can cause accidental invocation on loosely related prompts, exposing container, image, or compose operations without sufficiently specific user intent.

Missing User Warnings

Medium
Confidence
92% confidence
Finding
The skill documents destructive commands such as force-removing containers and pruning resources without warning, dry-run guidance, or confirmation requirements. In an exec-capable agent, these commands can irreversibly delete containers and associated state, causing service disruption and data loss.

Missing User Warnings

Medium
Confidence
88% confidence
Finding
The documentation includes image push and private registry login flows but omits warnings about credential handling, registry trust, and unintended data exfiltration. In an agent setting, this could lead to accidental publication of proprietary images or exposure of authentication secrets.

Sudo/Root Execution

Medium
Category
Privilege Escalation
Content
# 依赖说明
# 参考 docs.docker.com 安装最新版
# 2. 安装Compose V2插件
sudo apt install docker-compose-plugin  # Debian/Ubuntu
# 3. 验证V2
docker compose version
Confidence
81% confidence
Finding
sudo

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Static analysis

No suspicious patterns detected.