The Agent Skills Directory

[COMMAND_EXECUTION]: The skill attempts to list files in a hidden directory within the user's home folder (~/.claude/design-systems/). Accessing hidden directories in the home folder is a sensitive operation that can expose environment details or user configurations.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest data from external sources, including GitHub repositories, Figma links, and file attachments, which creates a surface for indirect prompt injection if those external sources contain malicious instructions.
Ingestion points: Scanning provided brief for GitHub/Figma links, images, and .md attachments.
Boundary markers: None identified in the instructions.
Capability inventory: File read/write/edit, directory creation, file copying, and browser devtools access.
Sanitization: No specific sanitization or validation of the external content is mentioned before it is used to influence the design process.
[DYNAMIC_EXECUTION]: The skill generates HTML artifacts containing React and JSX code, which are then served and executed locally. While this is the primary purpose of the skill, it represents the execution of dynamically generated code based on potentially untrusted input.

wireframe