The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill documents UI components that ingest and render external data (e.g., in searchable selects and custom option templates in references/patterns.md). This represents an inherent attack surface for indirect prompt injection if the processed data contains malicious instructions. However, the skill provides standard React/TSX implementation patterns and does not introduce specific vulnerabilities. Standard web development protections and LLM guardrails are expected to manage these runtime risks.
[SAFE]: No malicious patterns were detected. The skill consists entirely of documentation and code examples for a well-known UI library (Mantine). It contains no executable scripts, network operations, file system access, or obfuscated content.

mantine-combobox