yao-skill-manager

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.90). The SKILL.md's "SOP for External Skills" and related workflows explicitly instruct running commands like "npx skills add <repo_path>" and "npx skills add //external//" to download and inspect community/external skills (public repos), and the Gem Export Workflow requires extracting and synthesizing those skill contents so untrusted third‑party content can directly influence agent behavior.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The SOP instructs running "npx skills add //external//" (a GitHub/npm repository fetch) which at runtime downloads remote skill content (SKILL.md and code) that can define agent prompts or execute code, so this remote repo reference is a runtime dependency that controls agent behavior.