This is a follow-up to the article on security.txt. We have received a first report.
Someone wrote to us about a gap on our domain (a missing CAA record in the DNS). The message was courteous and clear. It is always welcome when someone tells you what is missing on your site.
The finding was accurate. The consequences described were overstated. Nothing showed access to our systems or to our data. I have not received another message since.
I fixed it. I was glad to. I also updated the free tool loi25.certi360.com so this point is checked from now on.
A security firm receives this kind of email too. If we receive it, you will too.
What is a CAA record?
A CAA record (Certification Authority Authorization) is a line in your DNS, the public directory of your domain. It says which certification authorities are allowed to issue an HTTPS certificate for your domain. Since September 2017, every authority must check it before issuing one.
Without a CAA record, any authority may issue a certificate for your domain. That is true. Someone who wanted to abuse that would still have to prove to that authority that they control the domain. If they can, you have a much bigger problem than a missing CAA record.
So this is one more layer of defense.
I added it.
Before you add a CAA record, list the authorities that issue your certificates today. We had three: Let’s Encrypt, Google Trust Services, and SSL.com (the last two through Cloudflare). If you forget one, the next renewal fails and your site shows an error page. The free site crt.sh lists the certificates issued for your domain. I wrote about this in the article on HTTPS certificate renewal.
The check is now in the tool
loi25.certi360.com now checks whether a CAA record is present. It is free. The report gave me a good reason to add it.
Bug bounty and beg bounty
A bug bounty program is a company inviting researchers to test its systems. There are rules, a defined scope, and rewards announced in advance, often through a platform such as HackerOne or Intigriti.
A beg bounty is something else. A free scanner runs across thousands of domains, then the same email goes out to everyone, with a request for payment. It only takes one company in a hundred paying for it to be worthwhile.
The expression comes from Chester Wisniewski and Michael Argast, at Sophos. Troy Hunt, the creator of Have I Been Pwned, made it widely known in 2021. Sophos reports demands of 150 to 2,000 US dollars per finding.
The email I received, I file on the useful, practical side. It did not ask for payment. Other messages ask to be paid. That is a beg bounty.
What a payment request often looks like
The template shows up often:
- a configuration finding, named in the subject line;
- your domain as the target;
- steps to reproduce, with a free online checker;
- consequences broader than the finding;
- a request for money, a bounty, or a paid service.
Often, nothing shows access to your data.
The findings that come back
They often come from free tools. The most common:
- no CAA record (that is the one I was told about);
- missing DMARC, or SPF judged “weak”: the records that stop a third party from sending email in your name;
- missing HTTP headers, such as X-Frame-Options, which stops another site from displaying yours inside an invisible frame to trap clicks (“clickjacking”);
- a software version visible in the server responses;
- a contact form with no limit on the number of submissions.
What they share: configuration findings. Not an intrusion. I am glad to receive them. They are often simple fixes, and they are worth doing.
Paying is a different matter
A courteous message that tells you about a flaw or a problem is welcome.
Paying for a scanner result funds the model and invites the next one. Sophos describes an organization that was asked for 500 dollars, then 5,000 dollars when more findings arrived.
Replying to a payment demand also confirms that a human reads the inbox. The next finding then has an address that answers.
The more dangerous variant is a fake report with an attachment, or a link to a “proof of concept.” That is phishing dressed up as disclosure.
How to recognize a disclosure that shows access
Real researchers exist. I do not want you to throw their emails out with the rest.
A disclosure that should go to the front of the queue describes a demonstrated effect on your systems or your data, not only a configuration gap. The person does not lead with money, or they point to your policy if you have one. They accept time to fix the issue. They often come through a known platform, or through the address in your security.txt file.
A configuration finding with overstated consequences can still be welcome. That is what happened here. I was glad to be told.
In practice
Before you act on an email like this, three questions:
- Is the tool cited a free online checker? You have a configuration finding. Useful, and often simple to fix.
- Is there proof of access to your data? If not, this is not an intrusion.
- Is there talk of a reward, a bounty, or a paid service? This is a beg bounty.
Then:
- Keep what is useful and fix it.
- Do not pay for a scanner result.
- Do not open an unexpected attachment, and do not click a proof link you were not expecting.
- Forward the email to your IT lead or your managed service provider. The finding deserves a look, even when the consequences are overstated.
- Publish a security.txt file, with a disclosure address and a clear policy. Ours says that we fix, that we can say thank you, and that no reward is paid. I explained how in the previous article.
In short, the finding is often true, and almost never urgent. I was glad someone wrote. At your organization, who receives these emails, and who knows how to keep what is useful?
Want to know where you stand? Let’s talk.
References