Security audit

Springboot Vue Builder

Security checks across malware telemetry and agentic risk

Overview

The skill is a coherent SpringBoot/Vue project builder, but it asks users to share high-value secrets in chat and has overly broad activation wording.

Review before installing. Use this only if you are comfortable with a project generator that orchestrates several subskills and creates runnable frontend/backend code. Do not provide real root database passwords or API keys in chat; use placeholders, environment variables, an untracked local secrets file, and a least-privilege database user instead.

SkillSpector

By NVIDIA

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (4)

Vague Triggers

Medium

Confidence: 91% confidence
Finding: The trigger guidance is overly broad and explicitly says the skill must be used whenever users mention common terms like '新项目', '全栈', 'SpringBoot', or 'Vue'. This can cause the agent to invoke a multi-step orchestration skill in unrelated contexts, increasing the chance of unintended data collection, unnecessary tool invocation, and context hijacking.

Vague Triggers

Medium

Confidence: 89% confidence
Finding: The skill description defines trigger scenarios in a broad, ambiguous way without clear boundaries or exclusions. In an agent environment, this kind of loose routing increases accidental activation and can funnel users into a workflow that requests secrets and spins up downstream skills they did not intend to use.

Missing User Warnings

High

Confidence: 98% confidence
Finding: The skill instructs the agent to ask users directly for a MySQL root password and DashScope API key in plain conversation, then reuse them in generated configuration. Collecting high-value secrets this way creates a strong risk of accidental logging, transcript retention, exposure to downstream tools, and insecure storage in files or prompts.

Ssd 3

High

Confidence: 99% confidence
Finding: This is a direct secret-handling flaw: the skill tells the agent to solicit and record highly sensitive credentials in plain language for later configuration use. In the context of an orchestration skill that passes artifacts across phases and subskills, this is especially dangerous because secrets may be copied into documents, code, config files, logs, or tool inputs, greatly expanding exposure.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Static analysis

No suspicious patterns detected.