The Agent Skills Directory

[SAFE]: The skill serves as an educational resource and architectural guide. It does not contain any executable scripts or instructions that would compromise the host system.
[IPC_SECURITY]: The instructions and examples demonstrate secure Inter-Process Communication (IPC) by using contextBridge to expose only a minimal, typed API surface to the renderer, preventing direct access to sensitive Electron internals.
[SANDBOXING]: The skill enforces secure defaults such as contextIsolation: true and nodeIntegration: false, which are essential for isolating untrusted renderer content from the main process and Node.js APIs.
[INPUT_VALIDATION]: Detailed examples are provided for sanitizing and validating all data received from the renderer in the main process, specifically addressing common vulnerabilities like path traversal and URL injection.
[NATIVE_API_SAFETY]: Guidance on using native APIs like shell.openExternal and dialog includes mandatory protocol validation and allowlisting to prevent arbitrary command execution.

desktop-framework-electron