The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to modify the project's task runner configuration (e.g., package.json, Makefile, pyproject.toml) to add execution commands for generated prototypes. This is a standard and expected behavior for a tool designed to facilitate rapid development and testing.
[DATA_EXFILTRATION]: While the skill mentions data persistence, it explicitly restricts it to in-memory state or local scratch files with clear temporary names. It advises against connecting to production databases, effectively mitigating risks of unintentional data exposure.
[PROMPT_INJECTION]: The skill implements an interactive surface for prototypes (keystrokes in TUIs and URL search parameters in web UIs). While these are ingestion points for untrusted data, the skill is designed for direct user interaction in a local development context with no evidence of complex capability chaining that would lead to injection vulnerabilities.

prototype