nlweb-data-loading

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

• Third-party content exposure detected (high risk: 0.85). High: the required runtime workflow for db_load.py fetches arbitrary outsider-authored web content (RSS/Atom feeds, JSON-LD URLs, sitemap-derived URLs, and URL-list files) and then parses the fetched body text into Schema.org objects that are embedded and stored for later LLM context.

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

• Potentially malicious external URL detected (high risk: 0.80). db_load.py explicitly fetches external URLs at runtime (e.g., the example feed URL https://example.com/feed.xml and arbitrary URLs from --url-list or sitemaps) and ingests their content verbatim into the retrieval/agent context, meaning remote content can directly influence agent responses.