notebooklm-researcher

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 untrusted public web content (e.g., via "nlm research start" which discovers web sources and "nlm source add --url 'https://...'" and then "nlm research import " / "nlm notebook query" to read and synthesize those sources), so third-party pages and user-generated content are read and used to drive generation and actions.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly instructs fetching external web content at runtime via commands like nlm source add --url "https://..." and nlm source add --url "https://youtube.com/..." which NotebookLM ingests into its context and therefore can directly control prompts/results.