SkillSpector

NVIDIA's security scanner for AI agent skills that detects vulnerabilities, malicious patterns, and security risks before an agent executes a skill. Acts as a guardrail layer that agents and skill registries can run to audit new tools or capabilities before loading them.

Connects to: AI agent skill validation, vulnerability detection, malicious pattern scanning, skill security audit · · Apache-2.0 8,903★

Use it with an AI agent

Loadbay is an MCP server, so an agent can search the catalog and find this harness:

claude mcp add --transport http loadbay https://loadbay.xyz/api/mcp

• Source: https://github.com/NVIDIA/SkillSpector
• This harness as JSON: /api/harnesses/skillspector
• Agent setup: /setup.md
• Browse all 700+ harnesses on Loadbay