The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides shell commands and aliases for project management and automated cleanup of development environments.
Evidence: The skill suggests a cleanup command find ~/prototypes -mtime +30 -type d -exec rm -rf {} \; and a corresponding shell alias proto-clean to recursively delete local directories older than a certain duration.
[DYNAMIC_EXECUTION]: The skill facilitates a workflow where executable UI code is dynamically generated via an LLM and written to the filesystem for immediate execution.
Evidence: The RapidIterator class contains a write_component method that saves generated React code directly to a file to trigger hot-reloading in a live development server.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted natural language descriptions to generate functional React components, creating a potential surface for instructions to influence the generated code.
Ingestion points: The flow_description parameter in StoryboardPrototype and the free-text description in the 10-Minute Prototype Protocol (SKILL.md).
Boundary markers: No specific delimiters or safety instructions are defined to separate user intent from the generated code boundaries.
Capability inventory: The skill utilizes file-writing capabilities (write_component) and triggers shell commands for project execution (npm start, npm run dev).
Sanitization: There is no evidence of automated sanitization or validation of the input strings before they are interpolated into the code generation prompts.

rapid-prototyping