The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes subprocess.run within scripts/run.py, scripts/setup_environment.py, and scripts/__init__.py. These operations are dedicated to environment management, including the creation of a Python virtual environment, installation of required packages via pip, and the automated installation of the Google Chrome browser.
[DYNAMIC_EXECUTION]: The scripts/run.py file serves as an orchestration wrapper that dynamically invokes other Python scripts in the scripts/ directory. This ensures that all skill components are executed using the correctly configured local virtual environment.
[INDIRECT_PROMPT_INJECTION]: The skill ingests text content from NotebookLM in scripts/ask_question.py. This data, derived from external documents, could contain malicious instructions. The skill lacks explicit boundary markers or sanitization for this ingested content. However, the agent's capabilities are limited to the provided scripts (file writes for library management and network access for browser automation), and the ingested text is not executed as code.
[EXTERNAL_DOWNLOADS]: During the setup phase, the skill downloads the Google Chrome browser via the patchright library. This is a required component for the browser automation functionality and is fetched from well-known official sources.
[DATA_EXPOSURE]: Authentication state and browser cookies are stored locally in the data/ directory to facilitate persistent sessions with Google services. The skill includes appropriate safeguards, such as .gitignore entries, to prevent this sensitive information from being accidentally shared or committed to version control.

notebooklm