SkillGuard scans untrusted skill content. Treat scanner bugs as security issues when they allow a malicious skill to bypass deterministic checks, crash the scanner, access host secrets, or trigger code execution.

Please report security issues privately through GitHub security advisories:

https://github.com/stella/skillguard/security/advisories/new

Do not include private skill contents, secrets, tokens, or production logs in public issues.

Until the first stable release, only the current main branch is supported.

SkillGuard aims to:

• avoid executing candidate skill files;
• avoid network access by default;
• avoid following symlinks;
• cap scan resources;
• produce deterministic reports.

SkillGuard does not guarantee that a scanned skill is safe to install.