A second opinion on any website.

SiteSentinel reviews a site the way a careful security analyst would, then records what it saw so you can check its work.

9
inspectors running in parallel, covering TLS and headers, phishing signals, DNS, WHOIS, links, speed, accessibility and SEO.
60+
individual checks. Each one records what it observed, the points it cost and how to fix it.
77
automated tests covering the scoring, the network safeguards and past false positives.

Why it exists

Most website scanners give you a letter grade and a list of red labels. That's fine until a result looks wrong, because there's nothing to check it against.

SiteSentinel shows its working. Each finding includes the header, HTML snippet or script it was based on, how many points it cost and why that matters. When a result could be a false positive, the report says so.

How the score works

Every category starts at 100. A failed check deducts its full severity value and a warning deducts half. No category drops below 15 from ordinary findings, so a single issue can't make a site look like a total failure.

The overall score is a weighted average, and security and safety count the most. Categories that couldn't run are left out, not scored as zero. If most of a site can't be reached, the report says Unable to Analyze and shows no rating. A confirmed malware or phishing match is the one case that drops the score to zero.

What it won't do

It only makes the requests a normal visit would: fetching public pages, reading response headers, looking up DNS and WHOIS records, and checking Google Safe Browsing when a key is configured.

It doesn't scan ports, try passwords, submit forms or send attack payloads. Before connecting, it refuses private network addresses, localhost and cloud metadata endpoints, and it checks again on every redirect so a hostile site can't point it somewhere internal.

How it's built

Read the source

SiteSentinel is an open-source project by Nicholas Darwin. Every rule and weight is in the repository, and issues and pull requests are welcome.

View on GitHub