SkillSpector

An AI skill security scanner that detects vulnerabilities, malicious patterns, and security risks.

Agent SkillDevOpen source

Type: Agent Skill
Open source: Yes
GitHub Stars: ★ 9.9k
Source: skill-github
Repository: github.com/NVIDIA/SkillSpector

Overview

SkillSpector is an AI skill security scanner designed to detect vulnerabilities, malicious patterns, and security risks before installing AI skills. It supports multiple input formats (such as Git repositories, URLs, zip files, etc.) and provides multi-stage analysis, including fast static analysis and optional LLM semantic evaluation. SkillSpector generates reports in various formats and offers risk scoring and recommendations. Suitable for developers and security teams to ensure the safety of AI skills.

Capabilities

▪Multi-format input scanning
▪Detection of 65 vulnerability patterns
▪Two-stage analysis: static + LLM semantic evaluation
▪Real-time vulnerability lookup
▪Multiple report output formats
▪Risk scoring and recommendations

Use cases

Security checks before AI skill installationSecurity scanning in continuous integration/continuous deployment (CI/CD)Enterprise internal AI skill security auditsDevelopers self-checking AI skill security

Setup

Requires: Python 3.12+API Key (LLM 分析时)

Install using uv or pip: `uv tool install git+https://github.com/NVIDIA/skillspector.git` or install from source: `git clone https://github.com/NVIDIA/skillspector.git && cd skillspector && make install`

This information was compiled by AI from public sources and may contain inaccuracies — please refer to the source.

FAQ

How to perform a quick installation?

Using the uv tool: `uv tool install git+https://github.com/NVIDIA/skillspector.git`

Is a Python environment required?

Yes, Python 3.12+ is required, but it can also be run using a Docker image.

Which LLM providers are supported?

Supports providers such as OpenAI, Anthropic, and NVIDIA Build.

Related skills

gemini-cli

Gemini CLI is an open-source AI agent that brings Gemini's powerful capabilities directly into the terminal.

MCP

private-gpt

Provides a complete API layer for local models, supporting RAG, skills, tools, and more.

MCP

planning-with-files

File-based persistent planning skill, ideal for AI coding agents and long-running tasks.

Agent Skill

scientific-agent-skills

Transform any AI agent into a scientific assistant with 147 ready-to-use research skills.

Agent Skill

susi_alexa_skill

A skill that enables question-and-answer interactions between Alexa and Susi AI.

Agent Skill

claude-context

Provides code search capabilities for Claude Code, turning the entire codebase into context.

MCP