The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill requires the notebooklm-py library from PyPI and recommends installing Playwright's Chromium browser. It also provides an automated installation workflow that fetches release tags directly from a public GitHub repository.
[COMMAND_EXECUTION]: The Python script scripts/nlm.py utilizes the subprocess module to execute the notebooklm CLI for authentication. Additionally, the skill documentation provides instructions for users to set up persistence through cron jobs or scheduled tasks to automate session refreshing.
[DATA_EXFILTRATION]: The skill manages sensitive Google session cookies, which are stored in and retrieved from a local JSON file (~/.notebooklm/storage_state.json). While necessary for the skill's functionality, this data represents a potential exposure risk.
[PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection.
Ingestion points: The skill ingests untrusted data from URLs, local files (PDF, markdown, docx), and raw text inputs.
Boundary markers: There are no specific delimiters or instructional guardrails used to isolate ingested data from the agent's logic.
Capability inventory: The skill has the ability to execute shell commands and perform network operations.
Sanitization: The skill does not perform validation or sanitization on content retrieved from external sources before processing it through NotebookLM.

notebooklm