pyautogui-automation
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The
ensure_dependenciesfunction inscripts/utils.pyautomatically downloads and installs thepyautoguiandpillowpackages from the official Python Package Index (PyPI) using pip if they are not already available on the system. - [COMMAND_EXECUTION]: The skill uses
subprocess.check_callinscripts/utils.pyto execute the system's Python interpreter and run pip installation commands for dependency management. - [INDIRECT_PROMPT_INJECTION]: The skill exposes high-privilege UI automation capabilities that could be exploited if the agent processes malicious data from an untrusted external source (e.g., a website or document) and uses that data as input for automation commands like
type_textorclick. - Ingestion points: Data supplied to parameters such as
--textinscripts/automation.py, which is subsequently passed to thepyautogui.typewritefunction inscripts/keyboard.py. - Boundary markers: No delimiters or instructions are used to distinguish between legitimate user commands and potentially malicious instructions embedded in processed data.
- Capability inventory: The skill possesses full control over user input devices (mouse and keyboard), can capture the entire screen, and can interact with active windows on Windows systems via the
pywin32library. - Sanitization: Input text is passed directly to the automation library without validation or sanitization, allowing for the potential injection of special keys or unintended control sequences.
Audit Metadata