# SkillSpector

> Security scanner for agent skills

Detects prompt injection, data exfiltration and malicious patterns in skills.

- Source: https://github.com/NVIDIA/SkillSpector
- Homepage: https://docs.nvidia.com/skills/scanning-agent-skills
- License: Apache-2.0
- Language: Python
- Stars: 19130
- Forks: 1665
- Contributors: 88
- Last commit: 2026-10-02
- Latest release: v2.12.0 (2026-09-23)
- Purpose: Agent skills, Security & identity
- Runs on: CLI
- For: Personal, Enterprise

## Worth score: 82/100

How much DigGitHub recommends it, from activity, adoption, docs, license and security signals.

- Popularity: 21/25
- Momentum: 10/20
- Maintenance: 25/25
- Community: 11/15
- Readiness: 15/15

## Alternatives

- [Security Audit Skill](https://diggithub.com/cloudflare/security-audit-skill.md): Coding-agent skill for security audits
- [Graphify](https://diggithub.com/Graphify-Labs/graphify.md): Turn a codebase into a queryable knowledge graph
- [Geth](https://diggithub.com/ethereum/go-ethereum.md): Go implementation of Ethereum
- [Caveman](https://diggithub.com/JuliusBrussee/caveman.md): Cut coding agent tokens by making it talk less
- [Bitcoin Core](https://diggithub.com/bitcoin/bitcoin.md): Reference implementation of Bitcoin
- [Tailscale](https://diggithub.com/tailscale/tailscale.md): Easy WireGuard mesh VPN
- [ECC](https://diggithub.com/affaan-m/ECC.md): Skills, memory and guardrails that make coding agents work better
- [Trivy](https://diggithub.com/aquasecurity/trivy.md): Find vulnerabilities, misconfigurations and secrets

---

Source page: https://diggithub.com/NVIDIA/SkillSpector
Updated: 2026-10-03