read the rules: when the static html is not the rendered page

huntress ctf 2026, day 1 · coursestack welcome lane

tools used: ordinary browser developer tools; no custom artifact parser.

this item did not use the game master: a start for read-the-rules returned http 503. the prompt put the flag in the html of https://ctf.huntress.com/rules, invisible on screen, and the answer goes into the "read the rules" lesson field on coursestack.

the key insight: curl, view source and the hydrated dom are three different evidence surfaces.

artifact anatomy

the initial response was a roughly 7 kb next.js shell: _next/static script and style references, a __NEXT_DATA__ json block holding the tenant configuration (domain, brand colors, logo ids), and a sentry baggage meta tag (sentry-public_key, sentry-release, trace_id), but no flag. those hashes were public application metadata, not candidates. the page's content is produced by client-side javascript after hydration.

curl -s https://ctf.huntress.com/rules | grep -oE '__NEXT_DATA__|sentry-[a-z_]*' | sort -u

extraction method

  1. open the rules page normally and wait for hydration.
  2. search the elements panel, not view source, for the flag prefix or a hidden element.
  3. inspect nodes with display:none, off-screen positioning, zero opacity, or text inserted by a client component.
  4. submit the value on coursestack, not to the game master.

failed approaches and pivots

fetching the page with a plain http client returned only the shell, and no headless browser was available in the environment at the time, so the rendered page had to be read by a person in a browser.

lessons

where we got stuck

solve times

public scoreboard · day 1 finish: below top 10

  • 2h38msingle#124 of 735 solvers
  • #54team day finishof 358 teams that cleared it, in 3h04m

from the public scoreboard: 09:00 et release to our flag.

{generated by a clanker}