The Agent Skills Directory

[SAFE]: The skill defines a set of operational guidelines for creating throwaway prototypes. It emphasizes security best practices such as avoiding data persistence by default, isolating business logic from user interface code, and ensuring that prototype artifacts are clearly labeled and intended for eventual deletion. No malicious commands, obfuscation, or exfiltration patterns were detected.
[COMMAND_EXECUTION]: The skill instructs the agent to integrate prototype execution into the project's existing task runners (e.g., pnpm, npm, make, python). This is a standard part of the development workflow and is handled within the local project context.
[DYNAMIC_EXECUTION]: The primary function of the skill is to guide the agent in generating new code artifacts (TUIs or UI components) based on user requirements. The provided templates use safe coding patterns, such as conditional rendering based on specific URL parameters, rather than unsafe dynamic evaluation methods.
[INDIRECT_PROMPT_INJECTION]: The skill processes user prompts to define the scope and functionality of prototypes. To mitigate risks associated with untrusted requirements, the instructions mandate that logic be isolated into pure, portable modules and that the resulting prototypes avoid interaction with production databases or persistent storage unless explicitly requested.

prototype