notebooklm

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

• Third-party content exposure detected (high risk: 1.00). The skill programmatically opens and scrapes NotebookLM pages (e.g., ask_question.py and the SKILL.md "Smart Add" flow that queries https://notebooklm.google.com/notebook/...) to read user-uploaded / publicly shared NotebookLM notebooks and then uses those responses to drive follow-up queries and library-management actions, so untrusted third‑party content can materially influence the agent's behavior.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill's runtime automation (ask_question.py / BrowserFactory) navigates to and reads responses from the user's NotebookLM URL (https://notebooklm.google.com/...), and those remote responses are injected into the agent's output/context (directly controlling what the agent returns), so this external URL is a runtime dependency that controls agent prompts/answers.