The Agent Skills Directory

[COMMAND_EXECUTION]: The skill is centered around the execution of the gws binary on the host system. It provides instructions for the agent to use various subcommands and flags to interact with the Google Forms API.
[PROMPT_INJECTION]: The skill handles data from Google Forms, which serves as a vector for indirect prompt injection. Malicious text embedded in form titles, descriptions, or user responses could attempt to override agent instructions during data processing.
Ingestion points: Data retrieved from Google Forms via gws forms get and gws forms responses commands.
Boundary markers: There are no specified delimiters or instructions provided to separate external form data from the agent's internal prompt context.
Capability inventory: The skill allows for the creation, modification, and retrieval of Google Forms and their settings, providing a range of actions that could be targeted by an injection attack.
Sanitization: No sanitization or validation logic is mentioned for the content ingested from the forms.

gws-forms