The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill automatically downloads and installs browser binaries (Chrome/Chromium) and Python dependencies during the first-time setup phase using standard tools like pip and Patchright. This is a legitimate functional requirement for browser automation.
[COMMAND_EXECUTION]: The skill employs subprocess.run within its management scripts (run.py, setup_environment.py) to handle virtual environment creation, dependency installation, and the execution of its internal automation logic. While this uses the shell, it is restricted to the skill's own internal script paths.
[INDIRECT_PROMPT_INJECTION]: The skill processes responses directly from the Google NotebookLM interface. This creates a potential surface where malicious instructions embedded in a user's uploaded documents could influence the agent's behavior.
Ingestion points: Content is retrieved from the browser in scripts/ask_question.py using CSS selectors defined in scripts/config.py.
Boundary markers: The skill does not use specific delimiters or 'ignore' instructions when returning retrieved content to the agent.
Capability inventory: The skill has capabilities to execute shell commands via subprocess.run (managed by run.py) and perform network operations via browser automation.
Sanitization: There is no explicit sanitization or filtering of the text retrieved from the external notebook source.

notebooklm