Delete the backdoor code, with a commit message saying you have removed a security vulnerability. Watch them try to defend leaving the backdoor in the system (favorite excuse I've heard is 'we need it for troubleshooting!') - they can rollback the commit if they really want the backdoor in. Send out resume.
That's not being sneaky at all; it's in the commit logs and he can announce it. It's just shifting the burden of action back to the people who want the vulnerability. In theory, the same. Practically, very different.
It's obviously not just a regular commit and everyone involved would know that. The commit log is not a proper place for a thinly veiled ethics complaint.
If you think something is wrong, have the courage to say so directly, don't hide behind a commit message.
As both employee and employer, it's my view that all employees should what they see as the right thing by default.
Personally, I think the poster should make the change and then announce it. I'm not saying he should be cowardly. If anything, I think the cowardly choice is to futz around trying to get consensus an elusive around doing the responsible thing.
Since they claim they won't use it and he says it was probably a mistake, this sounds to me like exactly the correct response: Let them save face and "agree" with the change if they are so inclined. A high handed accusation of an ethics issue does not look supportable at this time. And it is likely the worst thing he could do.
Doesn't surprise me. In that circumstance, I'd quit via a letter to the shareholders.
That's what I did when I worked for a large nefarious killing machine provider when I had a tony stark moment, grew some balls and worked out what they were doing was utterly wrong.
I informed them that I was bound by British law as well which supercedes any corporate rules and contracts.
Devil's advocate: assuming the company already quietly install updates, and the backdoor is not secured worse than the auto-update mechanism, this does not really give them additional capabilities. (Also, the phone likely already has more serious vulnerabilities.)
Of course, this exchange does suggest bad things about the company's ethics and competence.
The backdoor being called "backdoor" by this engineer already implies that it is not nearly as secure as the auto-update mechanism: the content is not signed by the company, it can't be disabled by the user, et c. I see no reason not to trust OP's judgement.
> I have found out recently that the remote assistant software that we put in smartphone we sell can be activated by us without user approval.
Maybe I'm missing something, but it says nothing about what might possibly happen in case their remote assistant is remotely activated. Also it's unclear how large is their user base. Everyone kind of assumes serious implications, though.
IMO if this whole thing is true, it indicates that the company probably doesn't have good QA and development process in place. Otherwise either such a bug would not exist (most likely it was left for debugging purposes), or it would really be a product of an evil intent (and hidden from uninitiated developers).
Seriously, is there any significant piece of hardware on stores nowadays that doesn't have multiple backdoors from application level down to the very microchips? Like [1].
please speak up. you should go to anybody and everybody you can and make your concerns known. it's possible that the people above you don't even know, but in the event they do, you should probably inform local news or someone.
I'm not saying that you should try and lose your job, but you should make your disagreement known.
Although not tech specific, at the most recent AngelHack hackathon, someone created a service where you can call in your confessions and people can listen to them. http://confessioncall.com/
