explain-via-html

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

• Third-party content exposure detected (high risk: 0.85). The skill’s runtime workflow fetches a public reference gallery page and an example HTML file from https://thariqs.github.io/... (outsider-authored web content), and then shapes the output based on that example’s structure/interactivity—so the LLM is exposed to outsider-provided text/HTML via the fetched page/example.

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

• Potentially malicious external URL detected (high risk: 1.00). The skill explicitly instructs the agent at runtime to fetch and follow the gallery page https://thariqs.github.io/html-effectiveness/, so external content is required and directly controls how the agent constructs its prompts/output.