readme-grader

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

• Third-party content exposure detected (high risk: 1.00). The skill's SKILL.md and README explicitly require fetching and analyzing README.md from a provided GitHub repository URL, meaning the agent will read untrusted, user-generated third-party content (public GitHub repos) and use it to drive grading and recommendations.

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

• Potentially malicious external URL detected (high risk: 1.00). The skill explicitly fetches README.md from user-provided GitHub repository URLs at runtime (e.g., https://github.com/sunny0826/kubecm), and that fetched content is injected into the agent's context to directly control its evaluation prompts—meeting the criteria for high-risk external content.