The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: Fetches configuration and the latest release tag from the vendor's official GitHub repository to facilitate installation of the notebooklm-py package. It also includes a command notebooklm skill install for setting up additional skill components.
[PROMPT_INJECTION]: The skill ingests untrusted data from various external sources, which is a potential surface for indirect prompt injection attacks where malicious instructions in the data could influence agent behavior.
Ingestion points: Uses notebooklm source add to import content from URLs, YouTube, PDFs, and multimedia files, and notebooklm source add-research to ingest web research data into the notebook context.
Boundary markers: No explicit delimiters or system instructions to ignore embedded commands within the sources were identified in the operational guidelines.
Capability inventory: The skill possesses significant capabilities including filesystem writes via download, complex content generation via generate, and the ability to orchestrate background tasks using a subagent pattern.
Sanitization: No sanitization, filtering, or validation of the retrieved external content is described before it is indexed and processed by the AI models.

notebooklm