The Agent Skills Directory

[PROMPT_INJECTION]: Indirect prompt injection surface detected. The skill uses user-provided search terms to generate design guidelines that are then persisted to markdown files like MASTER.md. The skill instructions tell the agent to follow these files as a source of truth for subsequent design work, which could allow a malicious user to influence agent behavior if the generated content includes instructions designed to bypass agent constraints. * Ingestion points: User query arguments in scripts/search.py used for database lookups. * Boundary markers: Absent in the generated markdown files, meaning retrieved instructions are presented as direct directives. * Capability inventory: The skill has file system write capabilities in scripts/design_system.py to persist these generated instructions. * Sanitization: Basic punctuation removal is performed in scripts/core.py during tokenization, but this does not prevent semantic injection.

ui-ux-pro-max