clarvia-aeo-check

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

• Third-party content exposure detected (high risk: 0.80). The SKILL.md explicitly instructs scoring arbitrary tools by URL (e.g., "Score https://github.com/example/my-mcp-server" and "Try scanning by URL directly with aeo_score"), which requires fetching and interpreting public, user-generated web content (GitHub/public sites) that can influence agent decisions and tool use.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill's prerequisite instructs running "npx -y clarvia-mcp-server", which will fetch and execute remote npm package code at runtime and is a required external dependency (npx clarvia-mcp-server).