notebooklm

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

• Third-party content exposure detected (high risk: 0.90). The skill programmatically opens arbitrary NotebookLM notebook URLs and scrapes the NotebookLM UI for answers (see scripts/ask_question.py which calls page.goto(notebook_url) and extracts response selectors, plus the SKILL.md/README Smart Add that instructs querying user-provided NotebookLM links), meaning it ingests third‑party/user-uploaded content from NotebookLM and thus can be exposed to untrusted, user-generated content that could carry 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 programmatically opens and queries user-supplied NotebookLM pages at runtime (e.g. https://notebooklm.google.com/...), and the fetched NotebookLM responses are injected into the agent's output/context to drive follow-ups and final answers, so external content fetched at runtime can directly control prompts/behavior.