Did this properly for once. Same list, randomised into three groups of 400, same body copy, only the subject line changed.
Group A — question subject ("Quick question about your resources page")
Open 41%, reply 9.2%, placement 1.8%
Group B — direct subject ("Broken link on /resources")
Open 38%, reply 14.6%, placement 4.1%
Group C — compliment subject ("Loved your guide on X")
Open 46%, reply 4.8%, placement 0.9%
The compliment opens best and converts worst, which tells you everything about what people do after they open. Direct and specific wins. Stop writing "quick question".
Link builder. If it ranks, it is whitehat enough for the client.