The Agent Skills Directory

[DYNAMIC_EXECUTION]: The skill guides the agent in generating executable code, including standalone HTML files for business logic demos and framework-specific components (e.g., TSX) for UI variations. This is a core functionality of the skill as described in LOGIC.md and UI.md.
[COMMAND_EXECUTION]: Instructions specify that generated prototypes should be run using standard project task runners such as pnpm, python, or bun. The agent invokes these commands to facilitate the execution of the prototype as noted in SKILL.md.
[INDIRECT_PROMPT_INJECTION]: The skill uses user prompts to determine the prototyping strategy and generate the associated code, which represents a potential surface for adversarial instructions.
Ingestion points: User prompts are processed to identify whether a logic or UI branch is required and to define the prototype's specific behavior as detailed in SKILL.md.
Boundary markers: The instructions do not prescribe the use of delimiters or 'ignore embedded instructions' warnings for the data ingested during the generation process.
Capability inventory: The agent has the capability to write files to the project directory and execute them using the project's task runners (SKILL.md).
Sanitization: The skill does not define specific sanitization or validation routines for user-provided input prior to its use in code generation.

prototype