The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses subprocess.run in scripts/__init__.py, scripts/run.py, and scripts/setup_environment.py to manage its execution environment. These commands are limited to creating a Python virtual environment, upgrading pip, installing packages defined in requirements.txt, and installing the necessary browser binaries (Chrome/Chromium) for automation.
[EXTERNAL_DOWNLOADS]: The environment setup process downloads standard Python packages from PyPI and browser binaries from official sources via the patchright library. These are required for the skill's core functionality of web automation.
[CREDENTIALS_SAFE]: Google authentication is handled interactively. The auth_manager.py script opens a visible browser window where the user must manually log in. The resulting authentication state (cookies and profile data) is stored locally in the skill's data/ directory. No evidence was found of these credentials being exfiltrated or sent to any third-party servers other than Google.
[PROMPT_INJECTION]: The skill includes instructions in SKILL.md and ask_question.py (via the FOLLOW_UP_REMINDER constant) that guide the AI agent's behavior to ensure research thoroughness. These are functional guidelines for the agent and do not attempt to bypass safety filters or override system constraints in a malicious way.

notebooklm