I fiddled a little to see how it works.
When importing (the tiddlywiki subcommand is load
, which term is preferred in this context?) tiddlers from an encrypted single file wiki, I am expected to provide the password in command line, in plain text, like this: $ tiddlywiki mywiki --password 123 --load my_encrypted_wiki.html
. I am not happy that the password is echoed to screen in plain text as I type it. I am not happy that the plain text password is eventually recorded in a shell history file. $ tiddlywiki --help password
doesn’t tell me that it could read the pasword from an environment variable for example. And tiddlywiki won’t let me “cheat” by not providing a pasword at all, like this: $ tiddlywiki mywiki --password --load my_encrypted_wiki.html
hoping that I’ll get a no echo password prompt. What can be done here to boost my opsec happiness?
Once I import an encrypted single file wiki, make some changes and export it again as single file - it is saved in plain text, unless I explicitly set a password again right before saving. I won’t go as far as claiming this is wrong, but did I not correctly interpret the output of tiddlywiki --help password
, which claims that “this sets a password for subsequent crypto operations”? If I load an encrypted wiki into the Node.js version and I’m telling it the password anyway, shouldn’t it use it when saving again later? I’m ok with manually entering a password again before saving, but I’d expect this to be needed only if I wanted to change it and save with another password.