The Agent Skills Directory

[COMMAND_EXECUTION]: The skill includes several Python scripts (gws_doctor.py, auth_setup_guide.py, gws_recipe_runner.py, workspace_audit.py) that facilitate Google Workspace administration by executing the gws CLI tool. These scripts use subprocess.run() with shell=False to execute commands from a local catalog or to check system status, which is the intended functionality of the skill.
[EXTERNAL_DOWNLOADS]: The skill instructs the user to install the gws CLI tool from the googleworkspace/cli organization on GitHub. This is an official and established tool for Workspace management, and the skill provides instructions for installation via standard, well-known package managers including npm, Homebrew, and Cargo.
[INDIRECT_PROMPT_INJECTION]: The skill processes potentially untrusted data from Google Workspace (such as the bodies of incoming Gmail messages or content within Drive documents) which is passed into the agent's context for analysis.
Ingestion points: Output from gws commands retrieving message lists, file metadata, and document content, which are then processed by output_analyzer.py or directly by the agent.
Boundary markers: Absent; the skill does not explicitly instruct the agent to use delimiters or ignore instructions embedded within the processed Workspace data.
Capability inventory: The skill provides a comprehensive command surface for Google Workspace administration, including the ability to send emails (gws gmail +send), delete files (gws drive files delete), and modify sharing permissions (gws drive permissions create).
Sanitization: Absent; output_analyzer.py performs field selection and aggregation but does not sanitize the values within those fields to prevent instruction injection.

google-workspace-cli