The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill allows the agent to ingest untrusted data from Google Workspace services and perform high-privilege actions based on that data.
Ingestion points: In SKILL.md, the agent is instructed to read email bodies (gog gmail messages search --include-body), document content (gog docs cat), spreadsheet data (gog sheets get), and chat messages (gog chat messages list).
Boundary markers: The instructions do not define boundary markers or explicit instructions to the agent to ignore instructions found within the retrieved content.
Capability inventory: The agent has wide-ranging capabilities across multiple files and scripts, including sending emails (gog gmail send), creating calendar events (gog calendar create), uploading files (gog drive upload), and sending chat messages (gog chat dm send).
Sanitization: No sanitization or validation mechanisms are described for the external content before it is processed.
[COMMAND_EXECUTION]: The skill is centered around the execution of the gog command-line tool in a shell environment.
Evidence: SKILL.md contains numerous examples of shell commands involving the gog utility and environment variable exports used to interact with Gmail, Calendar, Drive, and other Workspace services.

gogcli