Exactly what they would do, yes.
I'd add in general that almost all of the monitoring they do is electronic. They don't need to assign special agents to read every post here 24/7. They have algorithms that provide them exactly what they need at both the individual and collective levels.
That's not to say they don't have plants though. In doing so, they wouldn't need anywhere near 10% of the registry to be effective. They only would need to plant a few 'influencers', similar to those who have been known to exist already on other social media platforms. Unlike some mainstream reddit groups or twitter, where they are primarily trying to steer opinions or "re-educate", I think they'd be more interested in baitposting or contacting individual members here who are showing signs of mental instability or youngcels because they are more likely to be impressionable/manipulated like you stated.
I'd be more worried about them making a "useful idiot" or scapegoat out of someone than anything. In addition to setting people up to directly 'go off' as first person actors, these scumbags have also been known to pose as benevolents and then 'recruit' unsuspecting individuals who conform to a pre-existing 'profile' for a to-be-committed-false flag, the purpose of which is to vilify the group the recruit belongs to. After preparing some circumstantial evidence, they'll essentially "stage" the person in the locale and time the event is set to take place, and then scapegoat the recruit for whatever the crime was even though they had nothing to do with what happened.
Make no mistake though, they definitely do study us. Steve Bannon even said as such regarding the last US presidential election. We're labeled as a hate group by the SPLC for crying out loud. I have an island in the Pacific to sell anyone who honestly thinks that any group of people on that list doesn't get monitored closely by the government. We're on the radar of every relevant three letter agency in the US.