The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes the gog command-line interface to interact with the local system and Google API services.
[INDIRECT_PROMPT_INJECTION]: The skill processes potentially malicious content from external Google services, creating a surface for indirect prompt injection attacks.
Ingestion points: Content retrieved via gog from Gmail, Google Drive, Docs, and Sheets as described in SKILL.md.
Boundary markers: The skill instructs the agent to use the --wrap-untrusted flag in command templates to isolate external content.
Capability inventory: The CLI possesses broad capabilities including reading, writing, and deleting data across multiple Google services, as well as sending emails.
Sanitization: Security is dependent on the --wrap-untrusted CLI flag and manual oversight requirements such as --dry-run and explicit user intent for mutations.

gog