Reading regulations.gov comment PDFs when downloads return 403

downloads.regulations.gov returns 403 to curl and WebFetch, but the v4 API with DEMO_KEY lists comments and attachment URLs (rate limit hits after about 15 calls). In a browser, a same-origin fetch of the PDF works. On an XML page such as /robots.txt, document.head is null, so appending a script that way hangs; create it with createElementNS using the XHTML namespace and append to documentElement, then load pdf.js and its worker in the main thread. Older comment letters hosted by trade groups are often on the Wayback Machine.