What it does
SkillSpector statically inspects skill packages and reports suspicious instructions, vulnerable patterns, and security findings.
NVIDIA SkillSpector
NVIDIA's scanner for detecting vulnerabilities, malicious patterns, and risky behavior in AI agent skills.
Last reviewed: 2026-07-19
NVIDIA/SkillSpector
Repository facts
Strengthen testing, debugging, code review, security checks, and delivery verification.
- GitHub Stars
- 13.4K
- License
- Apache-2.0
- Default branch
- main
- Last pushed
- 2026-07-14
- Repository status
- Active
- Snapshot
- 2026-07-19
Best fit
Teams reviewing third-party skills before installing or distributing them.
Capabilities
- Scan skill files for prompt injection, unsafe commands, secrets, and suspicious behavior.
- Produce findings that can be reviewed before a skill enters the toolchain.
Use cases
- Audit a downloaded skill repository before enabling it for developers.
- Add a security gate to an internal skill publishing workflow.
Source
NVIDIA/SkillSpector
Open GitHub repoSource & review
Review notes: Details checked against the public repository README and project links.
Related in this category
Strengthen testing, debugging, code review, security checks, and delivery verification.
Playwright MCP
Microsoft's Playwright MCP server for browser automation through structured accessibility snapshots and deterministic tools.
Codex Plugin for Claude Code
OpenAI's plugin for invoking Codex from Claude Code to review code, delegate work, rescue a task, or transfer context.
Cybersecurity Skills
A large community collection of structured cybersecurity skills mapped to major security and AI-risk frameworks.