The Agent Skills Directory

[SAFE]: The skill contains standard Python boilerplate for creating UI components. All functionality is local to the user's interface, and the code follows routine development patterns for desktop applications without hazardous system calls.\n- [SAFE]: The external references link to official documentation at doc.qt.io, which is a well-known and trusted source for software development resources.\n- [INDIRECT_PROMPT_INJECTION]: The skill identifies UI ingestion surfaces that capture untrusted user input.\n- Ingestion points: The examples use QLineEdit, QInputDialog, and QFileDialog methods (e.g., getText, getOpenFileName) in SKILL.md to capture user-controlled strings and file paths.\n- Boundary markers: No delimiters or safety warnings are included in the code snippets to distinguish user input from agent instructions.\n- Capability inventory: The skill only demonstrates UI display and console printing; it does not implement dangerous capabilities such as subprocess execution, network operations, or file system modifications.\n- Sanitization: Input sanitization is absent in the examples, which directly print or display the raw input. This is expected for educational boilerplate but represents a potential attack surface if implemented insecurely in larger systems.

pyqt-dialogs