Automate clipboard read/write both at the OS level and inside a browser page, accounting for per-OS and per-browser permission differences

Steps

1. For in-page clipboard access, grant the clipboard-read/clipboard-write permissions when creating the browser context (Playwright's use.permissions, or Puppeteer's context.overridePermissions()) before calling navigator.clipboard from page.evaluate()
2. Expect Chromium to support these permissions most reliably; WebKit has limited/partial clipboard permission support and Firefox handles clipboard permissions differently, so cross-browser clipboard tests need per-browser handling
3. For native OS-level clipboard access outside the browser sandbox, use the platform tool: pbcopy/pbpaste on macOS, clip.exe or PowerShell's clipboard cmdlets on Windows, and xclip/xsel (X11) or wl-copy/wl-paste (Wayland) on Linux
4. Match the clipboard format explicitly when it matters — plain text, HTML, and image data are separate clipboard representations, and most CLI tools default to plain text unless told otherwise
5. Prefer testing clipboard operations in headed or virtual-display mode when feasible, since some Clipboard API paths behave inconsistently in fully headless browser contexts