notebooklm

CRITICAL E005: Suspicious download URL detected in skill instructions.

• Suspicious download URL detected (high risk: 0.80). The list includes a GitHub link to an unfamiliar/suspicious username (Kaliiiiiiiiii-Vinyzu) pointing to a "patchright-python" anti-detection fork (used to evade detection), which is a high-risk indicator for potentially malicious or tampered tooling.

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

• Third-party content exposure detected (medium risk: 0.65). The skill’s required runtime flow can pull outsider-authored free text from the public NotebookLM page/answer DOM (e.g., scripts/ask_question.py reads latest.inner_text() from notebooklm.google.com and appends it into the output used as the agent’s context).