59
React2Shell: maximum severity vulnerability discovered in React (www.tenable.com)
submitted 11 months ago by to c/programming@programming.dev
7 comments fedilink hide all child comments
 

I got an email from Vercel urging to upgrade Next.js based project 3 days ago. POC was published 2 days ago. Today I've checked my logs and I could already see attack attempts.

all 7 comments

sorted by: hot top controversial new old
[–] 4 points 11 months ago (6 children)

Fuck, Element for Matrix is apparently build on React, and I was updating like 4 days ago after few months.

Well, time to update again, I hope it's fine. Never really learned how to properly compromise-check your server.

  • permalink
  • fedilink
  • source
  • hideshow 6 child comments
    •
  • [–] 7 points 11 months ago* (1 child)

    Well, Element seems to still be running at the unupdated version even after update, so I'm just shutting the server down.

    I'm bummed that it took me 5 days to learn about it, does anyone have some tips how to get early warnings for techs you're using? I'm guessing there's a way with npm.

    Also, anyone has some tips how to properly compromise-check your server? I'm guessing there are logs to check for compromise, and audit your startup scripts for persistence? Any tools that could help with that?

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 1 child comment
    •
  • [–] 3 points 11 months ago* (3 children)

    it looks like this only applies react server components, and it doesn’t look like element uses react server components

    but i only had a quick skim; could be wrong, but personally i wouldn’t shut it down - not that im running a server myself

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 3 child comments
    •
  • [–] 2 points 11 months ago (2 children)

    I have no experience with React, so I couldn't tell. Thanks for the info, I'll keep it in mind.

    I think I've seen it mentioned that in case RSC isn't used, it might be vulnerable but it's not really confirmed, but you're right that it probably doesn't warrant shutting down the server.

    I don't really need it that much, though, so I'll just wait for the update, take a scour through logs and use it as a learning opportunity for forensics, and skip the reinstall.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 2 child comments
    •
  • [–] 2 points 11 months ago* (1 child)

    the vuln afaik is for remote code execution via basically a mechanism that’s kinda like a transparent RPC to the server (think like you just write frontend code with like a “getUsers” and it just automatically retrieves and deserializes the results so you can render the UI without worrying about how that data exists in the browser)

    i’m not a front end engineer, and haven’t used react server components, but i am a principal software engineer, i do react for personal projects, and have written react professionally

    i can’t think of a way it’d be exploitable via purely client-side means

    i THINK what they mean is that you can use some of the RSC stuff without the RPC-style interfaces, and in that case they say the server component is still vulnerable, but you still need react things running on your server

    a huge majority of react code is client-side only, with server-side code written in other languages/frameworks and interfaces with something like REST or GraphQL (or even RPC of course)

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 1 child comment
    •
  • [–] 1 point 11 months ago

    Any website using CSR only can't have a RCE because the code runs on the client. Any code capable of RSC that runs server and client side may be vulnerable.

    From what I've seen, the exploit is a special request from a client that functionally lets you exec anything you want (via Function's constructor). If your server is unpatched and recognizes the request, it may be (likely is) vulnerable.

    I'm sure we'll get more details over time and tools to manually check if a site is compromised.

  • permalink
  • fedilink
  • source
  • parent
    • 8 comments in c/programming@programming.dev view that post
    [–] 19 points 11 months ago

    Yeah been scrambling to get this one patched.

  • permalink
  • fedilink
  • source
    •
    [–] 11 points 11 months ago (5 children)

    This only affects sites that use the fancy new-fangled SSR rendering from what I've read. ReactJS (Client side) sites wouldn't be affected.

  • permalink
  • fedilink
  • source
  • hideshow 5 child comments
    •
  • [–] 8 points 11 months ago (3 children)

    Are we still vulnerable if our app doesn’t use React Server Functions endpoints?

    Potentially. According to the React Team, even if React Server Functions are not in-use, the vulnerability is still exploitable if React Server Components are supported.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 3 child comments
    •
  • [–] 6 points 11 months ago (2 children)

    I think their point was that CSR-only sites would be unaffected, which should be true. Exploiting it on a static site, for example, couldn't be RCE because the untrusted code is only being executed on the client side (and therefore is not remote).

    Now, most people use, or at least are recommended to use, SSR/RSC these days. Many frameworks make SSR enabled by default. But using raw React with no Next.js, react-router, etc. to create a client-side only site does likely protect you from this vulnerability.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 2 child comments
    •
  • [–] 4 points 11 months ago (1 child)

    Now, most people use, or at least are recommended to use, SSR/RSC these days.

    Why though?

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 1 child comment
    •
  • [–] 2 points 11 months ago

    I'm not the one recommending it lol.

    If I had to guess, it's to improve page performance by prerendering as much as possible, but I find it overkill and prefer to just prerender as much of the page as I can at build time and do CSR for the rest, though this doesn't work if you have dynamic routes or some kind of server-side logic (good for blogs and such though).

  • permalink
  • fedilink
  • source
  • parent
    • load more comments (1 reply)
    [–] 7 points 11 months ago

    We were very lucky that our usage was on the literal version before the affected version. Dodged a bullet.

  • permalink
  • fedilink
  • source
    •