A number of papers about a subliminal channel in the Digital Signature Algorithm were published more than ten years ago, allowing for communication through digital signatures in an undetectable manner. The subliminal channel is generally viewed as a method of legitimate but hidden communication, but it can also be used for leaking secret information (such as keys) in a undetectable way to anyone who knows what to look for. I will present on how this subliminal channel works, and demonstrate using a patched version of the GNU Privacy Guard how to use it for both benign and malicious reasons, both of which have little to no prior implementation in encryption programs.