The Agent Skills Directory

[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection and path traversal because it interpolates user-controlled variables directly into filesystem paths for output storage.
Ingestion points: The variables initiative-name and feature-name are derived from requester input to define the storage location (initiatives/[initiative-name]/design/) and filenames.
Boundary markers: There are no explicit delimiters or instructions provided to the agent to treat requester input as untrusted data or to ignore embedded instructions.
Capability inventory: The skill possesses file-write capabilities across the local directory structure defined in the Output section.
Sanitization: The instructions do not define any validation, sanitization, or restriction logic to prevent directory traversal characters (e.g., ../) or to ensure the input conforms to expected naming conventions.

create-design-brief