(solved) Nessus vulnerability false positives
-
I would suggest contacting Nessus as this issue is related to their software and the way its detecting pfSense. As Johnpoz have shown, the issue doesn't seem to be occurring to others.
-
@ johnpoz
Thanks for your work on this.
-
When I get back from my walk and snow blowing the drive - freaking lots of snow in chicagoland last night… I will fire up fresh 2.4.2 download on vm and see if can duplicate.. But I am unable to get it to show what your showing unless I do scan an OLD pfsense...
-
Hi,
That would be great. Last night I created a VM directly from the developer image and implemented it with the default setup… and I still got the ominous results. I used a fresh install of the community edition for Nessus and customer feedback is restricted to those who can afford the Pro License (~ $2200/yr).
The CGI vulnerabilities are not identified from the WAN side. The "unknown version" detection is almost certainly a false positive. If it can't be reproduced, then I am doing something (very) stupid.
-
Yeah I don't have the pro version either… do you have any sort of proxy or anything between your scanner and the pfsense lan IP other than switch? Just so we do apples to apples are you scanning via IP or fqdn?
I have some real life work to do ;) But will for sure spin up a fresh 2.4.2 vm. I am running scanner on a 16.04 ubuntu server VM..
-
Hi,
Yea, this work stuff always gets in the way of fun.
I have nothing unusual for my setups… no proxy, etc.
My native network is totally vanilla. A pfsense router and an unmanaged switch.
The VM networks consist of multiple VBox machines sharing an internal adaptor. I have two of these, one where the router is the stable release and another with the development snapshot from yesterday.
I have the Nessus community edition installed in Kali and, separately, in Arch Linux.
BTW: I am very impressed with pfSense and I will probably deploy it at the lab where I work..
-
Ok – so very odd... I just started a scan on the fresh vm.. 2.4.2 not p1 and it is showing same issues with the 2.1.1 errors..
now here is the thing... I set the web gui to be just on 80... While my main sg4860 is only on SSL... Let me change the 2.4.2 vm web to be on ssl only and rescan.
-
Well that wasn't it… Still getting errors with 2.4.2 not p1 using 8443...
URL : https://192.168.9.45:8443/
Installed version : unknown..0
Fixed version : 2.1.1Let me update it to P1.. and scan again.
edit: Ok so while the VM was updating to 2.4.2p1 I rescanned my sg4860... And not seeing the errors... So if this is clean after the update.. My GUESS would be that your system failed in its update to 2.4.2p1?? Give me a few minutes scanning the vm now.
edit2: Well WTF... So why is it clean VM shows the problem, but my sg4860 does not?? Could it really be something different in the CE version over the netgate version?? Going to have to look to the test they do against pfsense when it fails and then run that specific check against the sg4860... The only thing off the top I can think of is I am running a valid cert vs self signed even though nessus doesn't trust it. And am running ntop on 3000 as another webserver maybe that is confusing nessus.. Let me turn that off and scan m sg4860 again.
Ok this makes ZERO sense... When I seach the audit trail for this plugin ID shows pfsense not found on 80???
-
Well it uses pfsense_webui_detect.nbin in the nasl – this is clearly broken it seems...
-
Setting my web port to 8083 seems to correct the problem.
Perhaps a Nessus Pro subscriber could ring their bell on this. For 2200 bucks I say they should have some pull.
Meanwhile johnpoz, you really do an outstanding job of serving the community.
(No snow in Boston)
-
Great work johnpoz!
-
Thanks ivor but setting the gui to different port doesn't really fix anything - it just masks the problem. For whatever reason it seems that the nessus detection of pfsense is just broken.. I tried running the nbin that nasl script calls doesn't seem to output anything. I would have to dig way deeper than feel like doing ;)
They don't even seem to have a forum for other home users of the FREE activation can discuss problems and tricks, etc. Unless there is some 3rd party place which I have not looked into.. To be honest any such scan from the lan side kind of pointless if you ask me..
You should know without some scan telling you that your not uptodate… Everything else it told me like my snmp community was public, and it didn't trust the CA that signed the cert.. Oh you mean I allow snoop to unbound in the acl.. All stuff that already knew - the only little tidbit that was any sort of surprise was that the ntopng gui on 3000 was still using ssl 3, etc. I would be a bit concerned with that if it wasn't only access from my private secure network ;)
If you do get any more info MaxBishop I would be curious on their broken detection binary..
-
For what it's worth, I believe it's a benefit that a scanner is unable to properly determine what you're running. Why make it any easier on someone or something to figure out what you've got? :-)
-
Came across this because I'm having the same results w/ the newest version of Nessus and the newest version of pfSense. Did anyone ever get around to making a support ticket with Nessus? If we haven't gotten a response from someone with Nessus Pro, we might as well create one from a Nessus Free account. Better than nothing.
jimp, just because the current Nessus scanner doesn't detect the version doesn't mean it isn't possible. If the reason they can't fix it is because it isn't possible, that's another thing.
-
I don't think any of us has the (very) expensive Pro license. As best I can tell, there is no way to feed back to Tenable without one.
-
For what it's worth, I believe it's a benefit that a scanner is unable to properly determine what you're running. Why make it any easier on someone or something to figure out what you've got? :-)
Obscurity is not security. This is a bad line of thinking, especially if you wish to sell to Enterprises. Sure, hiding as much as possible from external attackers is nice but hiding from your CS department (or yourself) is generally not a good practice.
I came here as I too have the same problem on several Netgate boxes running 2.4.2_p1.
I'm not sure why this thread is marked as solved, it doesn't seem to be. I'll try and enquire with my support desk and see if I can get some answers about how the binary is detecting (or not, as is the case) the version. I don't know if they will have any real motivation to help, as I am low in the food chain and pfSense is not on the supported list. If I find anything helpful, I will report back. I am running with SecurityCenter, so I don't have as much control over the scans as you guys appear to.
I will also be trying a credentialed scan hopefully tomorrow and see if that changes things at all.
-
It's marked solved because it's not a pfSense issue. It's related to the way Nessus detects pfSense. If you want it fixed, please contact Nessus. Thread locked.
-
Obscurity is not security. This is a bad line of thinking, especially if you wish to sell to Enterprises. Sure, hiding as much as possible from external attackers is nice but hiding from your CS department (or yourself) is generally not a good practice.
This is not security by obscurity. It's reducing unnecessary information exposure. If you rely on the device itself to tell you what version something is, you need to have a proper mechanism setup and in place to do that internally (e.g. SNMP or other means of querying the device).
Being able to determine the OS based on network behavior or daemon responses is not a reliable detection mechanism, and being able to do so is a problem, not a solution. I wouldn't go so far as to say it's a security issue if you can identify the OS, but it's still better if it's not accurately discernible.