Element, area, full-page, and iframe captures

Element capture records a resilient DOM fingerprint and exact box. Area capture covers a freeform rectangle when no single element represents the feedback. Full-page capture scrolls and stitches the document. Iframe capture preserves frame boundaries and offsets.

• Prefer an element when the agent must identify code ownership precisely.
• Prefer an area for visual relationships across several elements.

Click a node, or press Enter on the current outline, to open an element pin. Drag a rectangle of at least six pixels to open an area pin instead. The first press on an iframe or frame element is ignored so the document inside that frame can take the selection.

Element pins record a fingerprint, a selector, and a DOM path that joins ancestor frames with a frame-boundary delimiter. Area pins store the rectangle and a pixel-size label without a locator. The copied screenshot still tiles around the union of every pin, including pins placed in child frames.

• The capture toolbar stays on the top frame; child frames show markers and the comment composer only.
• If a parent frame does not reply with its path, the pin keeps the inner document path only.
• Fixed or sticky elements are marked viewport-anchored so reopen does not treat them as document-scrolled boxes.