The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes subprocess.run within several management scripts (run.py, setup_environment.py, and scripts/__init__.py) to automate the creation of a Python virtual environment, install necessary libraries, and coordinate the execution of its internal Python toolset. These commands are localized to the skill's own directory.
[EXTERNAL_DOWNLOADS]: Upon first execution, the skill fetches external dependencies (patchright, python-dotenv) from the Python Package Index (PyPI). It also triggers a download of Google Chrome or Chromium browser binaries through the patchright CLI to facilitate the required browser automation.
[INDIRECT_PROMPT_INJECTION]: The skill acts as a bridge for unstructured data from Google NotebookLM, which introduces a vulnerability surface for indirect prompt injection.
Ingestion points: Text is retrieved from the browser's DOM elements via .inner_text() calls in scripts/ask_question.py.
Boundary markers: The skill does not currently implement delimiters or explicit 'ignore instructions' warnings when passing retrieved data back to the agent.
Capability inventory: The skill possesses capabilities for shell execution (via run.py) and file-system modifications (managing its internal database and authentication files).
Sanitization: The ingested content is not subjected to filtering or sanitization before being provided to the agent.

notebooklm