acquiring-skills

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

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly runs git clone of external repositories (https://github.com/anthropics/skills and https://github.com/letta-ai/skills) at runtime to fetch skill content that can include Markdown prompts and executable scripts which the agent may invoke, so the fetched content can directly control prompts or execute code.