The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: Fetches the official Google Cloud SDK installer from sdk.cloud.google.com during the setup process.
[REMOTE_CODE_EXECUTION]: Executes the Google Cloud SDK installation script as part of the recommended setup flow.
[COMMAND_EXECUTION]: The auth-setup.sh script automates dependency installation via pip and manages the local storage of authentication tokens in the user's configuration directory.
[DATA_EXFILTRATION]: The skill enables extensive access to sensitive user data within Google Workspace, including Gmail, Drive, and Docs, which is necessary for its automation functions.
[PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection due to its ability to ingest and process content from external sources.
Ingestion points: Data enters the agent context through API calls that read Gmail messages, Drive files, and Google Docs or Sheets content.
Boundary markers: No specific instructions are provided to help the agent distinguish between its core logic and potentially malicious instructions embedded in retrieved data.
Capability inventory: The agent can perform significant actions such as sending emails, modifying files, and managing calendar events based on processed information.
Sanitization: There is no evidence of data sanitization or validation performed on the text retrieved from Google APIs before it is evaluated by the agent.

google-workspace