My favourite part of these tools is the zany use of numbered file descriptors. `keypair` outputs the public key on fd 5 and secret key on fd 9. But signing reads the secret key on fd 8, while verification reads the public key on fd 4! Why aren't they the same?? I have to read the manpage every time.
That's such a user-hostile design decision. I can't fathom what justifies it (other than kinky taste).
Makes your commands unreadable without a manual, leaves a lot of room for errors that are quietly ignored. And forces you into using a shell that comes with its own set of gotchas, bash is not known to be a particularly good tool for security.
And to those who stay this adds flexibility: it doesn't. Those file descriptors are available under/dev/fd on linux, with named options you can do --pk /dev/fd/5. Or make a named pipe.
> Those file descriptors are available under/dev/fd on linux, with named options you can do --pk /dev/fd/5.
If you have a procfs mounted at /proc and the open syscall to use on it, sure (and even then, it’s wasteful and adds unnecessary failure paths). Even argument parsing is yet more code to audit.
I'd have otherwise guessed that this tool mainly exists just to test lib25519. Personally I'd only ever want a library, or some higher-level tool. A CLI tool that just does raw signing feels like a weird (and footgun-shaped) middle ground.
This mostly exists to test lib25519 and ostensibly to build systems with shell scripts (though: few people would do that). It is a weird and footgun-shaped middle ground.
If you just want a raw ed25519 private key then `head -c32 /dev/urandom` does the job. But usually you want a DER/PEM wrapper or similar, which the openssl cli tools handle nicely.
It's why no one has succeeded in replacing GPG: you need a lot of systems to work in order to have an actual viable one, the ability to spit out signatures from keys is required but not sufficient.
I was wondering the same thing. My best guess is that is to guard against operator misuse. Like usb-a only plugging in one way. Anything that is secret will never accidentally print to stdout. String interpolation in bash with `—option $empty` might be safer than `8<$empty`. Have to explore more but yeah, this is a new pattern for me as well.
Another possible factor driving the decision to use numbered file descriptors: the logic to validate that a file exists (or can exist) at a given path, is readable/writable, etc. gets punted to the shell instead of being something the program itself has to worry about.
Are you confusing the open openSSL library with the CLI? Absolutely none of this is true when used as a signing tool on the CLI. Seems like you just needed to rant, rather than answer my question. Which is fine: I do it to, but I was legit asking a question that you ignored and you seem to know about openSSL?
Sounds like the perfect place to embed credential stealing malware. Good thing they publish their code on an independent third-party public code sharing platform. Oh wait...
Short of suspecting a malicious tarball, I really can't think of a reason why "publish[ing] their code on an independent third-party public code sharing platform" would be a selling point. You're getting the source code straight from the horse's mouth this way.
(2024)
My favourite part of these tools is the zany use of numbered file descriptors. `keypair` outputs the public key on fd 5 and secret key on fd 9. But signing reads the secret key on fd 8, while verification reads the public key on fd 4! Why aren't they the same?? I have to read the manpage every time.
That's such a user-hostile design decision. I can't fathom what justifies it (other than kinky taste).
Makes your commands unreadable without a manual, leaves a lot of room for errors that are quietly ignored. And forces you into using a shell that comes with its own set of gotchas, bash is not known to be a particularly good tool for security.
And to those who stay this adds flexibility: it doesn't. Those file descriptors are available under/dev/fd on linux, with named options you can do --pk /dev/fd/5. Or make a named pipe.
it being option can be nice if you don't want your keys touching disk and need to pass it over to other apps.
it being default is insanity
> Those file descriptors are available under/dev/fd on linux, with named options you can do --pk /dev/fd/5.
If you have a procfs mounted at /proc and the open syscall to use on it, sure (and even then, it’s wasteful and adds unnecessary failure paths). Even argument parsing is yet more code to audit.
I think the design is pretty good as-is.
I'm curious, what do you actually use it for?
I'd have otherwise guessed that this tool mainly exists just to test lib25519. Personally I'd only ever want a library, or some higher-level tool. A CLI tool that just does raw signing feels like a weird (and footgun-shaped) middle ground.
This mostly exists to test lib25519 and ostensibly to build systems with shell scripts (though: few people would do that). It is a weird and footgun-shaped middle ground.
> I'm curious, what do you actually use it for?
FTA:
> These tools allow lib25519 to be easily used from shell scripts.
I've never used ed25519-cli, but not having to use a library is nice for someone who isn't a programmer.
The Venn diagram of "not a programmer" and "can safely use Ed25519" is two non-overlapping circles.
"this app needs me to generate a key and point to it in config" is plenty of overlap
If you just want a raw ed25519 private key then `head -c32 /dev/urandom` does the job. But usually you want a DER/PEM wrapper or similar, which the openssl cli tools handle nicely.
I don't consider myself a programmer and I can use Ed25519 safely. I do however understand computing fairly well.
I consider myself a programmer and ed25519-understander, but the idea of using it directly within a shell script terrifies me.
Simply combine this tool with `openssl enc` and your shell script is as secure as any shell script could be
Someone writing shell scripts is a programmer, for better or worse.
It's why no one has succeeded in replacing GPG: you need a lot of systems to work in order to have an actual viable one, the ability to spit out signatures from keys is required but not sufficient.
GPG is pervasive for the same reason git is pervasive: network effects. There are plenty of better alternatives.
Such as? I need an alternative which supports commutative trust relationships of some sort which are revocable.
Keybase or any of the tools inspired by keybase (foks.pub etc)
I was wondering the same thing. My best guess is that is to guard against operator misuse. Like usb-a only plugging in one way. Anything that is secret will never accidentally print to stdout. String interpolation in bash with `—option $empty` might be safer than `8<$empty`. Have to explore more but yeah, this is a new pattern for me as well.
Another possible factor driving the decision to use numbered file descriptors: the logic to validate that a file exists (or can exist) at a given path, is readable/writable, etc. gets punted to the shell instead of being something the program itself has to worry about.
Those descriptors like 5 could be mapped to anything, including descriptor 1, stdout.
What a strange convention. I'm partial to minisign, which works on plain old files.
This little CLI is not meaningfully an alternative for signify/minify. Here's a good piece on signify from its author (who also comments here):
https://www.openbsd.org/papers/bsdcan-signify.html
I’m guessing it’s to support the test framework it’s built with?
support is fine. Being default is crazy
[dead]
It's djb's web site so it's a djb design. With great genius comes great different thinking.
> It writes the public key to file descriptor 5, and then writes the secret key to file descriptor 9.
Is the project trying to compete with GPG for worst interface ? Magic numbers BAD, especially in something that will mostly be used in scripts
Why not zoidbe... I mean, why not open ssh? It's literally a CLI that does every crypto operation with every primitive (except some PQC)?
If you mean the OpenSSL CLI, it's hard to think of a more footgun-y cryptographic tool than the one that:
* defaults to unauthenticated encryption
* buries its one authenticated mode
* requires explicit command-line nonces
* defaults to an MD5 KDF
You could probably keep going for another 10 bullets. Never use the OpenSSL CLI for anything other than TLS stuff.
You can use ssh-keygen for signing and verifying signatures.
You can also use age[1] to encrypt payloads targeting ssh public keys. And decrypt using ssh private keys.
[1] <https://github.com/FiloSottile/age>
Yeah, the OpenSSL CLI sucks. So what's to be done?
Sure, we can build a 25519-specific tool with a less footgun-y interface. Fine, whatever, for that one use case.
Or we can build an alternative OpenSSL CLI that explodes OpenSSL and its numerous useful features in a general way and helps fix lots of use cases.
Nothing is to be done. Just don't use the OpenSSL CLI. It's a deeply cursed concept for a tool!
A command like cryptography swiss army knife useful though. If not openssl, then what?
It's useful as a toy and a learning tool, but for nothing else. For those two things, OpenSSL is fine as it is.
[flagged]
[dead]
Are you confusing the open openSSL library with the CLI? Absolutely none of this is true when used as a signing tool on the CLI. Seems like you just needed to rant, rather than answer my question. Which is fine: I do it to, but I was legit asking a question that you ignored and you seem to know about openSSL?
[flagged]
[flagged]
I can't find the source. Anyone can point to it?
The Download link in the header (https://lib25519.cr.yp.to/download.html).
Sounds like the perfect place to embed credential stealing malware. Good thing they publish their code on an independent third-party public code sharing platform. Oh wait...
Short of suspecting a malicious tarball, I really can't think of a reason why "publish[ing] their code on an independent third-party public code sharing platform" would be a selling point. You're getting the source code straight from the horse's mouth this way.
> feels like a weird (and footgun-shaped) middle ground.
hmm
> It is a weird and footgun-shaped middle ground.
Oh? HMMMMM :|