The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill defines an interface to read data from external sources (Google Form titles, descriptions, and user responses) through the get and responses methods. This data is ingested into the agent's context and could theoretically contain malicious instructions.
Ingestion points: Data enters the context via gws forms get and gws forms responses commands as described in SKILL.md.
Boundary markers: The provided documentation does not specify the use of delimiters or specific instructions to ignore embedded content in form data.
Capability inventory: The skill uses the gws binary for command execution, allowing the agent to read and modify Google Workspace resources.
Sanitization: There is no mention of sanitization or validation of the content retrieved from form responses.
[SAFE]: The skill functionality is well-defined and consistent with its stated purpose of managing Google Forms. No obfuscation, hardcoded credentials, or unauthorized network operations were detected. The reliance on the gws CLI tool is a standard pattern for productivity-oriented agent skills.

gws-forms