Guys, I have add MemberAccessPolicy to the API, which you can plug into a DefaultObjectWrapper (or to any BeansWrapper). I have also added WhitelistMemberAccessPolicy, to ease adding a restrictive policy. Please take a look. 2.3.30-SNAPSHOT is in the Apache snapshot repo, as usual. You can start out from here in API documentation: https://freemarker.apache.org/builds/fm2/api/freemarker/ext/beans/MemberAccessPolicy.html
So please review these, tell if you have any recommendations, or you see a way to circumvent this. (One risky thing I see is that we have a long deprecated default static instance of DefaultObjectWrapper, which if course doesn't use any custom MemberAccessPolicy. We use that static instance internally in FM2 on a lot of places. I will have to review all such cases, and also make it less probable that they can become exploitable later.) I will also create a new implementation for DefaultMemberAccessPolicy later. The current one does exactly what the old one did. The only real solution will be still WhitelistMemberAccessPolicy, if someone indeed doesn't trust the template authors. On Tue, Dec 24, 2019 at 6:31 PM Siegfried Goeschl < [email protected]> wrote: > HI Daniel, > > Would it make sense to come up with a separate chapter for the existing > FreeMarker documentation to explain the things in detail? > > Thanks in advance, > > Siegfried Goeschl > > PS: Last email for today - preparing Christmas dinner :-) > > > On 24.12.2019, at 18:23, Daniel Dekany <[email protected]> wrote: > > > > I responded to your mail that says "The problem described in the article > > was less about arbitrary people but someone who hacked through other > > security issues to become administrator with temple editing rights". So I > > thought that the premise there was that people who shouldn't be able to > > edit templates become able to do so. But it doesn't mater how it was in > > that case. Because as I said in the linked bug report, and as the FAQ > says, > > if you allow someone to edit templates with the default FreeMarker > > configuration, that's almost like if you allow them to edit Java files. > So > > whatever your application has right to do (like read the password file, > > launch missiles, etc.), the templates probably can do as well. The point > of > > discouraging complex/technical logic in templates (not just in > FreeMarker) > > was the MVC principle, where you should only put presentation logic into > > the templates. > > > > We can't provide a practically useful default configuration that's secure > > if you can't trust people that can edit templates, because the whitelist > > content is specific to the concrete application. By default ?new is not > > restricted (well, it can instantiate TemplatModel-s only, but that hardly > > saves anyone security wise). The reason ?api is still disabled by default > > is that if someone went through the pain of setting up FreeMarker to be > > safe(r) (which implies that you do not use the default ObjectWrapper, nor > > the default settings for ?new, and you are thoughtful with your > > TemplateLoader, as the FAQ says), then the new FreeMarker version where > > ?api was introduced should not open a new hole on your system. For almost > > all users though, ?api enabled by default would be better (it's mostly to > > allow users to work around TemplateHashModel limitations when dealing > with > > java.util.Map-s), but I have chosen the safer approach when I added it. > > > > The unsafeMethods mechanism will be updated, as things stand, despite > that > > it's not strictly backward compatible. It will be still a quite pointless > > mechanism. I don't know why was it added by the author (some 10-15 years > > ago, I think). > > > > On Tue, Dec 24, 2019 at 4:14 PM Siegfried Goeschl < > > [email protected]> wrote: > > > >> Hi Daniel, > >> > >> Not sure about your line of thoughts :-) > >> > >> My understanding > >> > >> * There is a recipe out there how someone can access the file system and > >> the setup was not bad security-wise - only "?api" built-in was enabled > >> * I think the "?api.class.getResource" and > >> "?api.class.getResourceAsStream" can be marked as unsafe method? > >> * I also think that ALLOWS_NOTHING_RESOLVER is not the default > >> configuration? > >> * I actually tried the published code and it reads my "/etc/passwd" :( > >> > >> If the assumptions above are correct - can this particular attack be > >> avoided? If so we should react and improve the configuration ... > >> > >> Thanks in advance, > >> > >> Siegfried Goeschl > >> > >> > >>> On 24.12.2019, at 11:50, Daniel Dekany <[email protected]> > wrote: > >>> > >>> The blog entry might starts its case with a privilege escalation > >>> independent of FreeMarker, but the question you got during your > >>> presentation wasn't about that, I think. But more importantly, some > real > >>> world applications do allow editing templates for users who aren't > >>> necessarily some kind of superusers. Right now, after they realized > that > >>> the problem exists at all, they will have to figure out a solution > >>> themselves. We are in a much better position to do the same. > >>> > >>> DOS-ing is certainly less of a concern in general, though unintentional > >>> DOS-ing (or I guess unintentional) was a problem for > >>> try.freemarker.apache.org in the past. My point there is just that if > >>> really everyone from the Internet can edit templates, then it will be a > >>> problem, I guess for any practical template language. > >>> > >>> > >>> On Mon, Dec 23, 2019 at 11:55 PM Siegfried Goeschl < > >>> [email protected]> wrote: > >>> > >>>> Hi Daniel, > >>>> > >>>> I guess I need to re-read the FreeMarker documentation and ticket - > >> having > >>>> said that > >>>> > >>>> * The problem described in the article was less about arbitrary people > >> but > >>>> someone who hacked through other security issues to become > administrator > >>>> with temple editing rights > >>>> * The people having that skills usually don't have any interest in > >>>> starting a DOS attack by messing up templates - there are more > valuable > >>>> things out there ... > >>>> * I think it is pretty much impossible to make FreeMarker 100% bullet > >>>> proof (tons of features, a lot of code, arbitrary libraries coming > from > >> the > >>>> application) but at least we can check that this attack does not work > >> any > >>>> longer > >>>> * From my understanding - usually there a couple of security > >>>> vulnerabilites leading to complete data breach :-) > >>>> > >>>> Thanks in advance, > >>>> > >>>> Siegfried Goeschl > >>>> > >>>> > >>>>> On 23.12.2019, at 22:30, Daniel Dekany <[email protected]> > >> wrote: > >>>>> > >>>>> Hi, > >>>>> > >>>>> In short, allowing untrusted users to edit templates is not > supported. > >>>> But, > >>>>> since people do it anyway, 2.3.30 will make an effort to allow doing > >> that > >>>>> with taking far less risk than what people take now. The > >>>> MemberAccessPolicy > >>>>> feature committed in recent days is the start of that. Actually, you > >>>> could > >>>>> always just use SimpleObjectWrapper (as the FAQ states), but clearly > >>>> that's > >>>>> too limiting for what many (most?) people use FreeMarker for. But > >>>> anyway, I > >>>>> don't believe that a template engine with the complexity of > FreeMarker > >>>> will > >>>>> be ever a good fit for applications where random people can edit > >>>> templates. > >>>>> If users are accountable in real life for what they did, like they > are > >>>>> employees at the client, then probably it will be good enough, but > not > >>>> for > >>>>> use cases where anyone can edit templates. If nothing else, you will > be > >>>> too > >>>>> easily DOS-able then. > >>>>> > >>>>> As of the blog entry, see this: > >>>>> https://issues.apache.org/jira/browse/FREEMARKER-124 > >>>>> Here I would add that it's likely that the calls used in the blog > entry > >>>>> won't work anymore in 2.3.30. I'm a bit uneasy about that, as it's a > >>>>> backward compatibility risk (it won't be just blocking that single > >>>> method), > >>>>> while it doesn't provide real security. You need a whitelist of > what's > >>>>> allowed for that (as opposed to a blacklist), and that's possible to > do > >>>>> with MemberAccessPolicy, but I will also provide an implementation to > >>>> help > >>>>> doing that. > >>>>> > >>>>> Also, in the FAQ: > >>>>> > >>>> > >> > https://freemarker.apache.org/docs/app_faq.html#faq_template_uploading_security > >>>>> > >>>>> > >>>>> On Mon, Dec 23, 2019 at 7:25 PM Siegfried Goeschl < > >>>>> [email protected]> wrote: > >>>>> > >>>>>> Hi folks, > >>>>>> > >>>>>> During my last presentation I was asked about how secure Apache > >>>> Freemarker > >>>>>> is in the context of user editing their templates - well, hard to > say > >>>>>> without knowing the application. > >>>>>> > >>>>>> But I came across an interesting article (see > >>>>>> https://ackcent.com/blog/in-depth-freemarker-template-injection/) > >> where > >>>>>> the authors successfully hacked a CMS based on Apache FreeMarker > >>>>>> > >>>>>> * As far as I know the UNRESTRICTED_RESOLVER is the default? Maybe > >>>>>> ALLOWS_NOTHING_RESOLVER would be a better default? > >>>>>> * Enabling "?api" needs to be enabled by developers which is fine > >>>>>> * Update the "unsafeMethods.properties" according to the article? > For > >>>> the > >>>>>> records "java.lang.Thread.suspend()" is duplicated anyway > >>>>>> > >>>>>> Thanks in advance, > >>>>>> > >>>>>> Siegfried Goeschl > >>>>>> > >>>>>> > >>>>> > >>>>> -- > >>>>> Best regards, > >>>>> Daniel Dekany > >>>> > >>>> > >>> > >>> -- > >>> Best regards, > >>> Daniel Dekany > >> > >> > > > > -- > > Best regards, > > Daniel Dekany > > -- Best regards, Daniel Dekany
