The Agent Skills Directory

[PROMPT_INJECTION]: The skill contains a surface for indirect prompt injection as it collects raw user input (e.g., product vision, feature descriptions) and embeds it into generated prompt files for implementation agents.
Ingestion points: Raw input is gathered via the /product-vision and /shape-section commands.
Boundary markers: The generated templates in references/export.md (e.g., one-shot-prompt.md) do not currently use explicit delimiters to isolate user-provided text from agent instructions.
Capability inventory: The skill performs file system write operations to create documentation and source code in the product/, src/, and product-plan/ directories.
Sanitization: No explicit validation or sanitization of user-provided content is performed before incorporation into exported artifacts.
[COMMAND_EXECUTION]: The skill manages project files and packages the output into a ZIP archive (product-plan.zip) during the export phase, which utilizes system-level file management utilities.

design-os