The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute the gws command-line tool with various flags to interact with Google Cloud's Model Armor service. It also suggests running gws generate-skills for environment setup.
[INDIRECT_PROMPT_INJECTION]: The skill serves as a surface for processing untrusted data by interpolating user-provided values into CLI commands.
Ingestion points: User-provided strings for --project, --location, --template-id, and --json flags within SKILL.md.
Boundary markers: The instructions include a [!CAUTION] block requiring the agent to confirm the command with the user before execution, which serves as a human-in-the-loop checkpoint.
Capability inventory: The skill is capable of executing shell commands via the gws binary, which performs network operations and file creation.
Sanitization: No explicit sanitization or escaping of the --json payload or other arguments is mentioned in the prompt instructions; safety relies on the underlying CLI tool's handling of inputs.

gws-modelarmor-create-template