Can a guard fake a round by copying a checkpoint tag?
Yes, and this is the deployment where the person reading the tag is the one with a reason to copy it. A plain tag returns the same bytes wherever it is, so a set of copies in a pocket produces a complete round without anyone walking it.
The reader is the attacker here
Almost everywhere else on this site, the person holding the phone is the one being protected: from a tag that was swapped, from a page that lies, from a terminal that is not what it claims. On a guard tour the roles are reversed. The tag is the employer's control and the phone belongs to the person the control is aimed at.
That changes which defences make sense. Tamper evident fixings and printed domains protect a member of the public from a substituted tag; they do nothing about somebody who owns the reader and wants a record that looks right.
Saying that plainly is worth more than a security feature, because a system sold as proof will be treated as proof by whoever relies on it.
How easy the copy actually is
Copying the content of a tag onto another tag takes a phone and a minute, which the copying page describes step by step.
Most tour systems therefore key on the serial number rather than the content, since a manufacturer serial cannot be rewritten on an ordinary tag, as the UID page explains. That raises the bar and does not close the door: tags whose serial can be chosen exist and are sold openly, which the fixed UID page covers.
Locking the checkpoint tag does not help either. It stops the tag being altered, not duplicated, and the locking page is explicit that the two are different properties.
Raising the cost
Cryptographic authentication is the only answer inside the tag. A chip that signs a challenge cannot be duplicated by copying what it says, because what it says is different every time, and the authentication page weighs the cost. A rolling code is the cheaper version of the same idea.
Everything else is corroboration outside the tag. A photograph at the point, a meter reading typed in, a question that has to be answered from where the checkpoint is, a satellite position recorded with the tap, a checkpoint whose position changes weekly.
And anomaly detection, which costs nothing and is what actually catches people: a round walked in four minutes, checkpoints always read in the same order at the same intervals, a whole tour recorded from one place. A copied round is regular in ways a real one is not.
What is proportionate
The question the worth cloning page asks applies here: what does a successful fake get somebody, and what does it cost the site.
For a round that manages work, deterrence and anomaly detection are enough, and the system's value was never in being unfalsifiable.
For a round that is a safety control, where a missed check has consequences, the tag is the wrong place to put the whole burden. Authentication in the chip, corroboration at the point, and a supervisor who occasionally walks it are three different layers, and the third is the one nobody can copy.
Frequently asked questions
Can a checkpoint tag be copied?
Its content trivially, and its serial number with tags sold for that purpose. Locking prevents alteration, not duplication.
Does using the serial number fix it?
It raises the cost, because an ordinary tag cannot have its serial rewritten. Tags with a settable serial exist and are sold openly.
What actually catches a faked round?
Patterns. A tour walked too fast, always in the same order, or recorded from one place looks different from a real one.
When is cryptographic authentication justified?
When a missed check has real consequences. For a round that manages work rather than safety, deterrence and anomaly detection are proportionate.