“Once it works out the wrinkles,” sure, but judging by the last 15 months, that could take a long time.
It’s about a year since the company CTO posted in these forums that they were going to make reliability a top priority, that they wanted SmartThings to be the last thing you thought of if you wondered why something didn’t work. Hasn’t gotten there yet. 
(Note the date on the following–one year ago. We should also note that the local processing that was promised in March 2015 is not what was delivered in October 2015. At the present time, very little can run locally and the “round tripping” described is still taking place one year later even for something as simple as turning on a Hue light at the same time every day.)
When it works well, SmartThings is my favorite home automation system. However, since last October I have yet to go 10 days without an impactful SmartThings failure.
I use a completely different system for home security, however, one for which I pay a monthly fee. Even when it’s working well, the SmartThings design just doesn’t need my own personal requirements for a security system. And my home security system has had only two false alarms in about 10 years. But everyone has different needs for peace of mind.
If you search the forums for security you’ll find lots of discussions. But if you just read the following thread, which includes comments from SmartThings engineers, you’ll find that there is a current platform instability issue leading to database corruption that they don’t have a fix for yet. I do expect it will get fixed eventually, but meanwhile you might want to look at one of the purpose-built security systems like piper, Scout, Abode, etc.
I would wait until SmartThings has been stable according to forum feedback for at least 90 days before considering it for a security system. It’s not just a question of whether they can fix a particular bug. It’s a question of whether the company really has committed to a philosophy of reliability first. And I think that can only be judged by actions, not promises or marketing materials. JMO