The Agent Skills Directory

[SAFE]: The skill operates as a structured interview and documentation generator. It follows a modular architecture defined by local protocol files and restricts its operations to the workspace context.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest data from user interviews to populate persona profiles. This represents a standard input surface for this type of tool. The instructions emphasize evidence-based documentation and the use of structured output (JSON and Markdown), which aligns with safe data handling practices.
[DYNAMIC_EXECUTION]: The skill describes the generation of an MCP (Model Context Protocol) surface and a single-page website. These outputs are constructed using local templates and predefined schemas based on a workspace protocol, rather than executing untrusted remote code.

customer-blueprint