mitm6 VS ql-https

I've been learning some common lisp, reading through Practical Common Lisp, and it's really neat. People say the good ideas of lisp got adapted in other languages and sure that's true of garbage collection, lambda's and some others, but I'm seeing plenty incredible stuff I haven't seen elsewhere, the condition system that among other things lets you fix and resume your program on exception, real interactive development, flexible object system, macros way more understandable than in other languages with AST macros as in lisp the AST is simple, an expressive dynamic language at high level of ruby and python while being an order of magnitude faster performance. Quicklisp also is really neat, how many other package managers can load new dependencies without restarting your application? And I was learning it with idea that it's not just of historical or hobby interest but legitimately a good choice I can use for new programming projects today for many tasks, but I just learned something that makes it impossible for me to consider, which is complete lack of security of quicklisp. You go to the website and see sha256 hash and PGP signature for quicklisp download, awesome it seems at the security standard you expect for a package manager. But then the actual quicklisp client does all downloads over http with no verification. What this means in practical terms is basically if you use quicklisp, anyone on your local network can easily hack your computer, by MITM (man-in-the-middle) the traffic and serving you backdoored software when you install packages from quicklisp. mitm6 will MITM windows machines on normal networks, bettercap can MITM linux and os x on most networks. Aside from attackers on your local network there's plenty other scenarios, you can go near office of CL using company and set up a open WIFI access point with same name as company wifi and hack their developers, using quicklisp over something like Tor is extremely dangerous at present as it would let the exit node backdoor the packages you download, and then in less likely but still should be protected against scenarios is just if quicklisp.org or any router between you and it is compromised, you can be hacked.

I'm seeing Responder work less and less often as more time goes by. There was a windows patch in 2016 that should result in Windows systems no longer trying to authenticate to hostnames resolved over LLMNR or NBT-NS. If you don't get any hashes with Responder, next try using mitm6 (https://github.com/dirkjanm/mitm6) or Pretender (https://github.com/RedTeamPentesting/pretender). If fully patched and properly configured, Windows hosts will only send credentials to hosts discovered via DNS, and since DHCPv6 is usually left unconfigured, you can poison DHCPv6 broadcasts to announce yourself as the preferred DNS server and you can get hashes or relay.

Solutions for the lack of https:

- add in https://github.com/rudolfochrist/ql-https (downloads packages with curl)

- use another package manager, CLPM: https://www.clpm.dev (or the newest ocicl)

> CLPM comes as a pre-built binary, supports HTTPS by default, supports installing multiple package versions, supports versioned systems, and more.

- use mitmproxy: https://hiphish.github.io/blog/2022/03/19/securing-quicklisp...

Other options are:

- Quicklisp -really slick, libraries in there are curated. (with https support here: https://github.com/rudolfochrist/ql-https and here: https://github.com/snmsts/quicklisp-https.git)

- for project-local dependencies like virtualenv: https://github.com/fukamachi/qlot

- a new, more traditional one: https://www.clpm.dev (CLPM comes as a pre-built binary, supports HTTPS by default, supports installing multiple package versions, supports versioned systems, and more)

For recent Quicklisp upgrades: http://ultralisp.org/

Ocicl is very new (5 days) and tries a new approach, building "on tools from the world of containers".

LPM's warning is not surprising. It's common for libraries (dare I say open-source ones?), even if they work well. It's part of the stability game, once they are marked 1.0, they are stable. LPM works well (as reported by others).

QL wants to do it portably, there are easy workarounds, but yeah…

(just saw https://github.com/rudolfochrist/ql-https)

That what I‘m doing: https://github.com/rudolfochrist/ql-https