This bot only reads.
What CyberCrushScanner does, what it does not do, and how to have your domain excluded.
How to recognise it
All our requests carry this User-Agent header:
Mozilla/5.0 (compatible; CyberCrushScanner/1.0; +https://cybercrush-v1.vercel.app/scanner)
Its behaviour
- Only GET and HEAD requests. Nothing that changes the state of your server.
- One request per second at most to the same domain.
- Forty requests at most per analysis, redirects included.
- robots.txt is read and respected for every path explored beyond the requested URL.
- No crawling: a handful of well-known public paths, from a fixed and public list.
- No intrusive testing: no injection, no fuzzing, no brute force, no authentication attempts.
40 requests at most · 1,000 ms between two requests · 10,000 ms at most per request · 5 redirects at most
Block the scanner yourself
You can refuse our requests from your robots.txt:
User-agent: CyberCrushScanner Disallow: /
Request exclusion of your domain
If you own a domain and do not want it to be analysable by our service, write to us at abuse@cybercrush-v1.vercel.app from an address on that domain, or from the contact published in its WHOIS. The domain and all its subdomains are added to our exclusion list, and any attempt to analyse it is refused without consuming a credit.
Logging
Every analysis is logged: the account that requested it, the target and the date. In the event of a complaint, we are able to answer.