Security audit

Web Dev Toolkit Free

Security checks across malware telemetry and agentic risk

Overview

This is a web-development guidance skill with local file and command capabilities that fit its purpose, though users should review any command or file-changing action before allowing it.

Install only if you want a Web development assistant that may read project files, suggest or make edits, and run local development commands. Review commands and file-changing actions first, especially anything involving deletion, reset, deployment tokens, or environment variables.

SkillSpector

By NVIDIA

Vulnerability Patterns

Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
Behavioral ASTexec() Call, eval() Call, Dynamic Import
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration

Findings (5)

Context-Inappropriate Capability

Medium

Confidence: 88% confidence
Finding: The skill exposes exec despite being described mostly as Markdown-based guidance, expanding its capability from advisory output into command execution on the user's system. In combination with broad trigger conditions and examples that normalize local command use, this increases the risk that an agent may run unnecessary or unsafe commands during ordinary development assistance.

Vague Triggers

Medium

Confidence: 84% confidence
Finding: The trigger conditions are broad enough to activate on common development conversations such as coding help, debugging, testing, or deployment. Overbroad activation is dangerous in an exec-enabled skill because it can cause the skill to engage in contexts where users expected only generic advice, increasing the chance of unintended file modifications or local command execution.

Vague Triggers

Medium

Confidence: 86% confidence
Finding: The keyword coverage list is excessively broad and ambiguous, making the skill easier to invoke unintentionally across many generic web or development prompts. In a skill with read/write/exec capability, an overly permissive activation surface materially raises the risk of accidental tool use and context-inappropriate actions.

Missing User Warnings

Medium

Confidence: 91% confidence
Finding: The skill describes create, modify, delete, export, save, import, and reset operations without warning about impact on user files or data. This is dangerous because it normalizes destructive or state-changing behaviors without consent boundaries, especially given the declared write and exec capabilities.

Missing User Warnings

Low

Confidence: 79% confidence
Finding: The command examples are simple and common, but presenting runnable local commands in an exec-enabled skill without any caution reduces user awareness that commands execute on their own machine. While the listed commands are low risk, the pattern can condition users or agents to treat future command execution as routine and safe without review.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.