OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
Agent Security & Sandboxing

Skillspector

Skillspector is an open-source Python security scanner for AI agent skills. It examines inputs for vulnerabilities, malicious patterns, prompt injection, data exfiltration and supply-chain risks before installation.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Skillspector — OpenViST project profile in Agent Security & Sandboxing.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use it before installing an unfamiliar skill, before publishing a submitted skill, or when adding skill checks to a development or supply-chain pipeline.

Who it is for

  • Developers evaluating agent skills before installation.
  • Teams reviewing skills submitted to a registry or publication pipeline.
  • Security engineers integrating skill analysis into CI or internal supply-chain workflows.
  • Users of Claude Code, Codex CLI, Gemini CLI and MCP skills.

Documented uses

  • Scan a skill directory, file, repository, URL or ZIP archive.
  • Detect documented vulnerability and malicious-behaviour patterns.
  • Generate machine-readable or human-readable scan reports.
  • Assign risk scores and suppress accepted findings during rescans.
  • Add a pre-publication or pre-installation security gate for agent skills.
Before you choose

Requirements & limits

The repository is public, open source and licensed under Apache-2.0. It is documented for CLI and self-hosted use, including installation from source, uv or pip, and a locally built Docker image.

  • The scanner supports multiple input types, including repositories, URLs, ZIP archives, directories and single files.
  • Static analysis is available, with optional LLM semantic evaluation as a second analysis stage.
  • Reports are available in terminal, JSON, Markdown and SARIF formats.
  • The analyzer documents 71 patterns across 17 security categories.
  • Remote and archive ingestion has a 100 MiB cap, ZIP archives have a 10,000-member cap, and individual analysis has a separate 1 MB file cap.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Install Skillspector from the Git repository

Set up the CLI for local agent-skill security scans.

Before you start

  • Python 3.12 or later.
  • uv or pip.
  • A virtual environment; the documented make targets assume it is active.

Steps

  1. Clone the NVIDIA/SkillSpector repository and change into its directory.

  2. Create and activate a virtual environment with uv or Python's venv.

  3. Run make install for production use, or make install-dev for development dependencies.

Expected outcome

The Skillspector CLI is installed in the activated environment and ready for documented scan commands.

Guide limitations

  • This recipe documents installation only, not a completed scan.
  • The project installs additional third-party open-source software whose license terms should be reviewed.
  • The documented quick-install path also supports uv tool install for a CLI-only installation.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
1 October 2026
Original website title
NVIDIA/SkillSpector · Official homepage title · retained in its source language
How we review sourcesSuggest a correction