The Agent Skills Directory

[COMMAND_EXECUTION]: The installation instructions include the use of sudo make install, which grants the installation process administrative privileges on the host system.
[EXTERNAL_DOWNLOADS]: The skill requires cloning and building software from an external GitHub repository (github.com/steipete/gogcli.git) and suggests installation via a third-party Homebrew tap.
[CREDENTIALS_UNSAFE]: The skill instructs the user to handle sensitive Google OAuth 2.0 client secrets and explicitly references storage in sensitive file paths like ~/Downloads/client_secret_....json and ~/.config/gog/.
[DATA_EXFILTRATION]: The tool provides extensive capabilities to read and transmit sensitive data across Google Workspace (Gmail, Drive, Sheets, etc.). This includes the ability to search emails, export sheets, and upload files to remote storage, which could be abused to exfiltrate data.
[PROMPT_INJECTION]: The skill processes untrusted data from Gmail, Drive, and other Google services, creating a risk surface for indirect prompt injection.
Ingestion points: The skill retrieves data via gmail search, drive ls, calendar events list, sheets list, and contacts search (SKILL.md).
Boundary markers: There are no instructions provided to the agent to treat retrieved content as untrusted or to ignore embedded instructions.
Capability inventory: The skill allows sending emails (gmail send), uploading files (drive upload), and modifying calendars/tasks.
Sanitization: There is no evidence of sanitization or validation of the content retrieved from external Google services before it is processed by the agent.

gogcli