theming-context

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

• Third-party content exposure detected (high risk: 0.70). The SKILL.md explicitly instructs the agent to "Use web_fetch or another method to retrieve the relevant documentation page" (e.g., docs.optics.rolemodel.design and the GitHub repo), so the agent will fetch and apply public third-party web content to its decisions, which could enable indirect prompt injection.

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

• Potentially malicious external URL detected (high risk: 1.00). The skill explicitly instructs using web_fetch at runtime to retrieve documentation pages (e.g. https://docs.optics.rolemodel.design/?path=/docs/components-button--docs), meaning remote content fetched from that URL can be injected into and directly influence agent responses.