• pixxelkick@lemmy.world
    link
    fedilink
    arrow-up
    1
    arrow-down
    2
    ·
    edit-2
    11 months ago

    That’s not broad enough.

    If you in any way have functionality that handles anything remotely requiring security, do not use HTMX.

    This goes way beyond “parameterized endpoints”.

    Listen extremely closely and pray to God anyone dev with more than 2 brain cells groks how serious th8s vulnerability is:

    HTMX enables arbitrary invocation of ANY api endpoint with cookies included, through html attributes, which inherently can’t be covered by Content Security Policy

    This is deeply important for any web dev worth their salt to understand.

    Sanitizing User input should be your LAST layer of defence against attack vectors. Not, NOT, your first and only

    It’s supposed to be your “break in case of emergency” system, not your primary (and only remaining) defense layer.

    • rwhitisissle@lemmy.world
      link
      fedilink
      arrow-up
      2
      ·
      edit-2
      11 months ago

      HTMX enables arbitrary invocation of ANY api endpoint with cookies included, through html attributes, which inherently can’t be covered by Content Security Policy

      I want you to please explain how HTMX bypasses the Content Security Policy connect-src directive, or any -src directive, for that matter, assuming it is specified (which it should be). Because I’m genuinely curious why the HTMX dev team would include a section on CSP in their docs if it did literally nothing, as you say.

      Actually, as an even more basic question…you do know that HTMX is literally just an AJAX library, right? It doesn’t actually “do” anything via HTML attributes. The additional HTMX attributes, like hx-get, hx-post, etc. just tells HTMX where and how to make the API requests. These requests are executed by the browser’s native fetch or XMLHttpRequest APIs, depending on compatibility and implementation. Therefore, HTMX is subject to the same security constraints and policies as any other JavaScript-based operation that makes HTTP requests. Which also, by definition, means that it adheres to the Content Security Policy directives configured for that website.

      In other words, an HTML button element with hx-get=“https://www.some-endpoint.com/” on it would eventually translate into

      const xhr = new XMLHttpRequest();
      xhr.open("GET", "https://www.some-endpoint.com/");
      xhr.send();
      

      on click.

      You do understand that, right?

      • pixxelkick@lemmy.world
        link
        fedilink
        arrow-up
        1
        arrow-down
        1
        ·
        edit-2
        11 months ago

        Just to be clear, are you talking about some kind of templating library that literally transpiles all the htmx logic and instead packs it into individual ajax logic in js files “per element”, such that you don’t need to serve htmx client side and instead you pre-transpile all the ajax logic out to separate files?

        Cause the very start of my statements was that if we had something like that then HTMX would be fine, as a templating lib that transpiled out to html+js.

        That you can CSP lockdown, because now you no longer are able to invoke arbitrary logic with html attributes, only the explicitly transpiled ajax can and all concepts of htmx have been actually removed from the final html+js you actually serve to the client.

        If that is what you are talking about above, then please link me because that sounds awesome and is what HTMX outta be, and would remove all of its security issues.

        If that’s not what you are talking about, and you truly dont understand the fact that you can’t compare an html element that triggers logic (which you can’t CSP block), to a script chunk that performs logic (which you can CSP block), then I think you do indeed need to go read up on and understand what the point if CSP is and why it was implemented in browsers.

        The two are apples and oranges. HTML elements should not be capable of invoking logic arbitrarily, that violates a core principle of html.

        • rwhitisissle@lemmy.world
          link
          fedilink
          arrow-up
          2
          arrow-down
          1
          ·
          edit-2
          11 months ago

          Just to be clear, are you talking about some kind of templating library that literally transpiles all the htmx logic and instead packs it into individual ajax logic in js files “per element”, such that you don’t need to serve htmx client side and instead you pre-transpile all the ajax logic out to separate files?

          My brother in Christ, what the fuck are you talking about “transpiling HTMX” and “serving HTMX client side?” You don’t “serve” HTMX and there’s nothing to “transpile into JavaScript.” It is JavaScript. That’s like saying you “serve React client side” and “transpile JavaScript into more JavaScript.” Jesus, I feel like I’m taking crazy pills.

          Cause the very start of my statements was that if we had something like that then HTMX would be fine, as a templating lib that transpiled out to html+js.

          Oh, okay, so you don’t actually know what HTMX is or how it works, then? Because HTMX (https://htmx.org/) is a JavaScript library. Like, literally just a JavaScript library. It’s like…4000 lines of JavaScript. In fact you can read the source code for it here: https://github.com/bigskysoftware/htmx/blob/master/src/htmx.js. For some…insane reason you seem to think HTMX is its own language. It’s not. It’s…just a JavaScript library. There is no other language called HTMX. There is no other mechanism or tool called HTMX. No implementation or protocol or ANYTHING else. It’s just a small JavaScript library.

          invoke arbitrary logic with html attributes

          Once again, HTMX enhances HTML with various attributes declaratively. It utilizes custom data attributes in HTML (like hx-get, hx-post) to specify how elements on the page should behave - essentially, how and where to fetch data or submit forms without a full page reload. This is a form of declarative programming that tells the htmx.js library (which is just doing fucking AJAX) what to do when certain events occur (e.g., a click or a form submission). The actions (like the actual requesting of data from an endpoint) are performed by the code in htmx.js.

          This is a fancy way of saying “if you stick an hx-get attribute on a button, then you can just say where you want a GET request to go to and what element you want updated with the HTML returned from it and htmx.js will parse that out on page load and set an event listener for the button click to know when to initiate an AJAX request to the defined endpoint.” If you had an hx-get attribute in an element in a page and that page didn’t have the htmx.js library loaded it would do literally nothing.

          And, once again, HTMX, being a JavaScript library, operates under the same security constraints as any JavaScript executed in the browser. This means that:

          1. HTMX’s scripts themselves must be loaded from sources allowed by the script-src CSP directive.
          2. Any dynamic requests to load content or submit data initiated by HTMX are subject to CSP’s connect-src directive.
          • pixxelkick@lemmy.world
            link
            fedilink
            arrow-up
            2
            arrow-down
            2
            ·
            11 months ago

            I see you don’t understand what the word “if” means, and you also don’t understand modern js practices.

            That’s like saying you “serve React client side” and “transpile JavaScript into more JavaScript.” Jesus, I feel like I’m taking crazy pills.

            You don’t serve react client side, any junior dev is familiar with transpiling framework code to produce their website. Yes, you 100% transpile react code before serving it, the fact you dont understand what I am talking about speaks volumes. It’s clear this whole time I’ve been having a discussion with someone who doesn’t even know the absolute bare minimum of day 1 front end dev. If you don’t understand how literally normal and industry standard something as basic as transpiling js is, you have literally zero business spreading info about something far more serious as HTMX.

            You are in zero way qualified to be recommending anyone expose their websites to the security nightmare that is HTMX, stop spreading misinfo, stop encouraging devs to do so.ething stupid, and go learn the basics of FE dev practices.

            If you don’t understand the tools of the trade, stop spreading terrible info about them online.

            Everything you have written in this entire thread has made everyone who has read it stupider and you have actively made the internet a worse place. You are a prime example of the exact thing that is wrong with web devs nowadays.

            Go back to the drawing board, you have a LOT to learn still it sounds like.