Find all anchor URLs (including in web components' shadow root nodes)

findAnchorURLs.js

function recursiveFindAnchors(node) {
    if (!node) return []

    // Find anchors URLs
    const bareAnchorURLs = [...node.querySelectorAll('a')].map(anchor => formatHref(anchor.href))

    // Find all shadow roots
    const allShadowRoots = [...node.querySelectorAll('*')].filter(node => node.shadowRoot).map(node => node.shadowRoot)

    // No web components in here
    if (allShadowRoots.length === 0) return bareAnchorURLs

    // Accumulate and go a deeper level in each, you can use flatMap() if supported instead
    return bareAnchorURLs.concat(allShadowRoots.map(recursiveFindAnchors).flat())
}

function formatHref (href) {
    if (!href) return ''
    const {origin, pathname} = new URL(href)
    return origin + pathname
}

console.log([...new Set(recursiveFindAnchors(document.body))])