brief-outline-generator

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

• Third-party content exposure detected (high risk: 0.95). The skill’s runtime workflow ingests outsider-authored free text by calling web_fetch(sitemap_url) and then dataforseo:on_page_content_parsing for each discovered URL, which returns fully rendered page text from arbitrary third-party sites into the agent’s LLM context.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly calls web_fetch at runtime on the user-supplied sitemap_url (e.g. https://firefly.ai/sitemap.xml) to extract site pages and build domain_context that is then used to steer prompt generation and outline content, so that external URL fetch directly controls the agent's prompts/content.