CMS Detection

Identify which CMS a website is using from public signals — name, evidence, version when it is exposed, and a confidence rating.

Warning: Authorised use only. Use this tool for authorised security-assessment recon only. You must have permission to test the target.

Install

Python 3.10 or newer. requests and beautifulsoup4 install with the package.

pip install lupaxa-cms-detection
cms-detection --help

CLI

cms-detection https://example.com
cms-detection https://example.com --active
cms-detection urls.txt --format json --output results.json
python -m lupaxa.cms_detection --version

Default mode uses the homepage plus a few well-known public files for CMS that already scored on that page. --active adds broader path probes, and when the homepage has no signal it also walks the catalogue's confirm files. A path is never a hit from HTTP 200 alone.

If the URL has no scheme, the tool tries https:// first and falls back to http:// only when TLS or connect fails.

Pass one URL, or a file of URLs with one URL per line. When the argument is an existing file, it is read as a list. Otherwise it is treated as a URL.

CLI Flags

--output writes the chosen format to a file. Results stay off stdout.

A text line looks like {url} => {cms} {version} ({confidence}). --verbose appends a short evidence note. JSON is a list of result objects. CSV columns are url, cms, version, confidence, evidence, candidates, and error.

Network and HTTP failures set error on that result and the batch continues. An invalid URL or empty input exits non-zero. Unknown CMS is a successful result: cms is empty and error is empty. A failed write of the output file is reported after the scan, and the CLI exits non-zero.

Library

from lupaxa.cms_detection import detect, detect_many

result = detect("https://example.com", active=False)
print(result.cms, result.version, result.confidence)

results = detect_many(
    ["https://example.com", "https://example.org"],
    workers=10,
    active=False,
)

detect takes one target. Keyword-only arguments are active, delay, retries, and timeout. detect_many runs those detects independently, keeps input order, and adds workers.

Network and HTTP failures do not raise. They become result.error with cms set to None. An invalid URL or empty input raises CmsDetectionError.

Results

DetectResult is a frozen dataclass.

Evidence fields are kind (meta_generator, header, cookie, script, path), value (what was seen), and cms (which signature it supported).

Confidence Rules

Each matching signal adds points. Those points pick the winner and are not shown in the CLI.

cms is the highest total. Other names with a score greater than zero go in candidates. Ties break on more distinct evidence kinds, then a version being present, then alphabetical name.

A path probe counts when the status is allowed and a required snippet appears in the body or headers.

Development

make init
make python-install-dev
make python-check