The Agent Skills Directory

[COMMAND_EXECUTION]: Path traversal vulnerability in scripts/design_system.py. The persist_design_system function constructs directory and file paths using the user-provided project_name and page arguments. These inputs are only modified by replacing spaces with hyphens and are not sanitized for path traversal sequences (e.g., ../), allowing an attacker to write files outside the intended design-system/ directory.
[PROMPT_INJECTION]: Indirect prompt injection surface. The skill ingests untrusted user input via the query and page_query arguments and interpolates these strings into markdown content formatted for the agent's context in scripts/design_system.py. The absence of clear boundary markers or sanitization for instructions within these strings could allow an attacker to influence the agent's behavior during the design planning phase.

ui-ux-pro-max