No. If you make and distribute security-related software, you should consider the safety of your user.

Your threat model absolutely should include this exact scenario. And you should know enough to understand and implement principles like plausible deniability and repudiation.