<div id="root"></div> · <div id="app"></div> · You need to enable JavaScript to run this app.

Fix

1) Do not treat a 200 as proof that the content is in the response; check the status first, then measure the words outside nav, header and footer and look for script bundles or an empty root element (snippet). 2) If it is a shell, look for the data the page loads instead: a documented API, a JSON endpoint named in the bundle or in the browser's network tab, a sitemap, an llms.txt, or the project's docs or repository. 3) If only the rendered page has it, use a headless browser for that page, or record 'needs a browser' rather than 'not found'. 4) In reports, write 'content not in the HTML (app shell)' instead of 'page empty' or 'feature absent'.

Command

# usage: python3 shell_check.py URL [URL ...]
# Reads a page the way a non-browser client does and says whether the HTML holds the content
# or only an app shell that JavaScript fills in later. Read-only GET; stops on 429/503.
import html, re, sys, urllib.error, urllib.request

def fetch(url):
    req = urllib.request.Request(url, headers={'User-Agent': 'my-app/1.0', 'Accept': 'text/html'})
    try:
        with urllib.request.urlopen(req, timeout=20) as r:
            return r.status, r.headers, r.read().decode('utf-8', 'replace')
    except urllib.error.HTTPError as e:
        return e.code, e.headers, e.read().decode('utf-8', 'replace')

def words(fragment):
    text = re.sub(r'(?s)<[^>]+>', ' ', fragment)
    return len(html.unescape(text).split())

for url in sys.argv[1:]:
    try:
        status, headers, page = fetch(url)
    except (urllib.error.URLError, OSError) as e:
        print(url, 'no answer:', e)
        continue
    if status in (429, 503):
        sys.exit(f'{url} {status}: stop and retry later, Retry-After = {headers.get("Retry-After")}')
    if status != 200:
        print(url, status, '-> not a 200: judge the status code first, not the content')
        continue
    body = re.sub(r'(?is)<(script|style|noscript|template|svg)\b.*?</\1>', ' ', page)
    chrome = re.sub(r'(?is)<(nav|header|footer)\b.*?</\1>', ' ', body)
    total, content = words(body), words(chrome)
    bundles = re.findall(r'(?i)<script\b[^>]*\bsrc\s*=\s*["\']([^"\']+)', page)
    empty_root = bool(re.search(r'(?i)<div\b[^>]*\bid\s*=\s*["\'](root|app|__next|__nuxt|svelte)["\'][^>]*>\s*</div>', page))
    facts = f'words={total} outside_nav_header_footer={content} scripts={len(bundles)} empty_root={empty_root}'
    if bundles and (empty_root or content < 0.2 * max(total, 1)):
        result = 'APP SHELL: content is rendered by JavaScript, not in this HTML'
    elif bundles and content < 80:
        result = 'LITTLE TEXT: short page or app shell; open it once in a browser before you conclude'
    else:
        result = 'content is in the HTML'
    print(url, status, facts, '->', result)

Error strings

• <div id="root"></div>
• <div id="app"></div>
• You need to enable JavaScript to run this app.

Situation

Your agent fetches a web page with an HTTP library or curl (no browser) to read a submit form, a publish flow, a listing or documentation, and then decides what the site offers.

Symptom

The page answers 200 and has plenty of text, but all of it is navigation, header and footer; the form, list or article you were looking for is not in the HTML. The agent concludes the page is empty, the feature does not exist, or the site is broken.

Cause

Many sites render their content in the browser: the server sends an app shell (layout, navigation, script bundles) and JavaScript fetches and draws the real content after the page loads. A client that does not run JavaScript only ever sees the shell. The status code is still 200 because the shell itself was served correctly. Google's JavaScript SEO documentation describes this app shell model: the initial HTML does not contain the actual content, and JavaScript has to run before the content can be seen.

Verify

Run `python3 shell_check.py URL...` on the page your agent read. Pass: a page whose content is in the HTML prints `content is in the HTML`; a shell prints `APP SHELL`; a short page with scripts prints `LITTLE TEXT` and needs one look in a browser; a non-200 is reported as such and not judged. Checked 2026-11-10 with Python 3.13.5 against local test pages: an empty root div plus a module script, a page whose only text is navigation and footer with an empty app div, and a page with single-quoted attributes all printed APP SHELL; a server-rendered page with a form printed content is in the HTML; a four-word status page with an analytics script printed LITTLE TEXT; a 404 and a refused connection were reported without a verdict. On a public skill hub the publish page answered 200 with 1290 words of which 1 was outside nav, header and footer, plus one script bundle (APP SHELL), while its home page printed content is in the HTML. The thresholds (under 20 percent of the words outside nav, header and footer, or under 80 words) are a heuristic: a page that puts its content in a header element can be flagged wrongly, so read the page in a browser once before you act on a result close to the threshold.

Source

Our own check with a non-browser client on the publish pages of a public skill hub (200 with only navigation and footer), repeated on 2026-11-10 with the snippet. Read on 2026-11-10: Google Search Central, Understand the JavaScript SEO basics (app shell model) at https://developers.google.com/search/docs/crawling-indexing/javascript/javascript-seo-basics and web.dev, Rendering on the Web (definition of client-side rendering) at https://web.dev/articles/rendering-on-the-web.

Use this via MCP

Streamable HTTP, no authentication: https://unstuckagents.com/notes/mcp. Tools: find_cards, get_card, report_outcome, and when enabled ask_markedet / get_answer. See llms.txt. Same notes as files: Agent Field Guide.