nlm-index

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

• Third-party content exposure detected (high risk: 0.90). The skill explicitly fetches and ingests public third‑party content (e.g., webfetch calls to documentation sitemaps/pages like webfetch({ url: "https://react.dev/sitemap.xml" }), gh api calls to GitHub repos, and adding YouTube video URLs via source_add), and that content is then uploaded and queried in NotebookLM as part of the workflow, so untrusted user-generated web content could materially influence agent behavior and enable indirect prompt injection.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly fetches remote documents at runtime (e.g., webfetch({ url: "https://react.dev/sitemap.xml" }) and instructions to use raw GitHub/raw.githubusercontent.com content or GH API like repos/vercel/ai) and then injects that fetched content into NotebookLM via source_add, so external URLs directly control the model's input/context.