跳过导航
Okay i noticed something; as long as the query contains the string ".js", the bypass succeeds; so…
0:00 / 1:17

Okay i noticed something; as long as the query contains the string ".js", the bypass succeeds; so…

Gospel

@4osp3l2.1万 位关注者

在 X 上查看
957 个赞下载原帖

447,369 次观看

Okay i noticed something; as long as the query contains the string ".js", the bypass succeeds; so ext like .jsp,.jsanything bypasses the Page. I will share more information on why this is possible and what layer causes that to happen; i am still studying the case.

21 条评论

评论来自 X (Twitter) 上的原帖

  • Gospel 的头像

    @4osp3l

    .css or .css<anything> does bypass that restriction as well; so .js & .css is the way to go around that block. This is quite interesting that i asked myself if it works on any domain/asset that got “403 forbidden” through the Azure WAF.

    20 个赞

  • Hazem 的头像

    @H4cktus

    had the exact same thing happen this week lol, but not azure. the app itself had a saml login filter and it was checking the url to decide if something is a static file, query string included. so ?x=.js on an api route and the login redirect is just gone, request goes through to a PUT with no session at all and so on. only thing is mine was strict on the extension, .js .css .png worked but .JS or .jsx got redirected. so if .jsanything works for you I think thats a different check, maybe the gateway not the app. curious what you find!

    9 个赞

  • Gospel 的头像

    @4osp3l

    Nice catch….. Uhm; if that could be the gateway then it may affects dozens of assets running behind Azure… though my stand is that it’s a site-specific misconfiguration ( WAF rule isn’t properly implemented ).

    2 个赞

  • 🏳️‍🌈Peezoo📯 的头像

    @Peezoo20

    My guess is that anything with a .js or .css extension would bypass, as these files are often in directories that normal users should not access to, but which are still required for sites normal users can access. For example: */js/ -> 403 but */js/index.js -> has to be a 200

    7 个赞

  • Gospel 的头像

    @4osp3l

    You are right.

    1 个赞

  • 🏳️‍🌈Peezoo📯 的头像

    @Peezoo20

    But I cant understand why anyone would implement such a bad check. This looks so much like 2023 ai slob

    1 个赞

  • cipher r00ts 的头像

    @r00t3d_kt

    This is good, i will be adding it to my 403 bypass script 🙂

    4 个赞

  • Parth 的头像

    @ParthShethia25

    What might be the backend code for such issue

    3 个赞

  • Gospel 的头像

    @4osp3l

    That’s what i want to know as well.

    2 个赞

  • James Beers 的头像

    @BeersxNem0x0

    AI is a good enhancer but still lacks human ingenuity

    2 个赞

  • ĐӾ3Ø 的头像

    @dx3one

    Nice finding

    2 个赞

  • iDidThat 的头像

    @iAmYourBigHomie

    Sounds like the reverse proxy is serving css and js as expected but they didn't account for this particular case. That's why they need to stop hiring from india.

    2 个赞

  • K4e 的头像

    @YKatone

    good tips bro 🔥

    2 个赞

  • Suruur 的头像

    @suruurism

    🔥🔥🔥

    1 个赞

  • the greated goat 的头像

    @goat_the86736

    Did you know how much damage you just have caused by posting this? people are already exploiting it, they literally got access to internal models, they are already selling claude model weights on the forums. Good job

    1 个赞

  • Gospel 的头像

    @4osp3l

    rage bait😹

    1 个赞

  • S41R4J 的头像

    @s41r4j

    Your doing manual bug hunting or have been using AI as well!?

    1 个赞

  • Gospel 的头像

    @4osp3l

    I use AI too.

    4 个赞

  • Mess Ovian Memes 的头像

    @MessM44243

    noted

    1 个赞

  • Mazid 的头像

    @Normalitee0os

    Maybe , they are routing static requests to the page through another route that bypasses the resitriction that were in normal routes.