The Agent Skills Directory

[SAFE]: No malicious patterns such as prompt injection, data exfiltration, obfuscation, or unauthorized command execution were detected.
[PROMPT_INJECTION]: The skill identifies an attack surface for indirect prompt injection as it processes untrusted user strings to generate files. However, it implements a closed vocabulary and validation logic that effectively maps input to specific design tokens, which serves as a sanitization mechanism.
[COMMAND_EXECUTION]: The skill correctly identifies the need for file_write permissions to generate its intended outputs, DESIGN.md and brief-preview.html.

design-brief