The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill provides instructions for installing the notebooklm-mcp-cli package from the official Python Package Index (PyPI). It emphasizes the importance of pinning specific versions for security.
[COMMAND_EXECUTION]: The skill includes local shell scripts (scripts/setup_notebooklm_mcp.sh and scripts/verify_notebooklm.sh) designed to automate the installation, configuration, and testing of the integration.
[DYNAMIC_EXECUTION]: The setup script (setup_notebooklm_mcp.sh) executes an inline Python snippet to programmatically update JSON configuration files, allowing it to support multiple AI agent platforms with varying configuration formats.
[INDIRECT_PROMPT_INJECTION]: The skill exposes a surface for indirect prompt injection.
Ingestion points: Data enters the system via the source_add tool which accepts external URLs and file paths (referenced in SKILL.md).
Boundary markers: The instructions do not define explicit boundary markers or delimiters for ingested content.
Capability inventory: The associated nlm tool can read/write files and perform network operations to Google domains.
Sanitization: There is no evidence of local sanitization or filtering of external content before processing.
[PERSISTENCE]: The provided setup script modifies configuration files for several AI agent platforms (such as Claude Code, Cursor, and Devin) to register the MCP server. This ensures the server is persistently available to the agent across sessions, which is the primary intended functionality of the skill.

notebooklm-mcp