CyberLock 9.0

  • Thread starter Thread starter danb
  • Start date Start date
  • Featured
sidenote possibility: After installing Cyberlock 9.00 beta, I went to bed and this morning my browsers in win10 host are not connecting, but I can ping ok? and browsers work fine in fedora_VM Guest OS. Last night I was not able to download InstallCyberLock900betaFirewallFix.exe but I did DL it today in fedora except it will not run, so must be corrupted somehow, currently uninstalled CL 9.00 beta and still host connection is snafu'd -- not saying CL is the issue, but I do think I need a nap, maybe it will fix itself when I wake up :ROFLMAO::sick::rolleyes::unsure:
 
I am super tired of talking about this. If we need to increase the daily tokens in the future, we certainly will. Looking at the data, 90-95% of users use less than 8000 tokens a day, and most are 0. I am in the same boat, I never install new software, so I never use tokens.
Less than 8000 tokens a day! From my experience, a single analysis can take over 10000—the highest I noticed was 18000! Anyway, I'm also tired of talking about this and will end the discussion here.
 
  • Like
Reactions: Zero Knowledge
@danb earlier I said I encountered an odd internet issue, and it seems like I identified an issue with DefenderUI 2.00 getting "locked up" in the windows platform at layer 7 or something like that per chatGPT. I tried to uninstall DUI but it failed to fully close or exit and finally I had to go into services and manually stop the service and then after a reboot (with DUI now uninstalled) the internet issues cleared up. I think I have seen / experienced something similar with previous DUI versions so might be unique to my system?
 
Hi Dan, I hope this isn't a dumb question but you talked about possibly adding a full system scan option in SiriusGPT / SiriusLLM / Cyberlock. Is this still an option or is it too cost ineffective. :unsure:
Thank you for reminding me... yeah, at some point we will probably add a smart scan that scans all of the common locations, and then eventually a full system scan. But it would probably be best to wait until the database cache is built up a lot more. Eventually we will have our own GPUs for Sirius analysis, but even then we have to limit compute costs somehow.
 
sidenote possibility: After installing Cyberlock 9.00 beta, I went to bed and this morning my browsers in win10 host are not connecting, but I can ping ok? and browsers work fine in fedora_VM Guest OS. Last night I was not able to download InstallCyberLock900betaFirewallFix.exe but I did DL it today in fedora except it will not run, so must be corrupted somehow, currently uninstalled CL 9.00 beta and still host connection is snafu'd -- not saying CL is the issue, but I do think I need a nap, maybe it will fix itself when I wake up :ROFLMAO::sick::rolleyes::unsure:
The original CyberLock 9.00 beta installer might have created some Windows Firewall rules, if you have that option enabled. The best thing to do is to either manually remove them, or install the latest version of CyberLock and go to the SiriusGPT tab in Settings and click the "Clear Firewall Rules" button at the bottom. I am certain the file is not corrupt on our end, otherwise other users would have the same issue.
 
@danb earlier I said I encountered an odd internet issue, and it seems like I identified an issue with DefenderUI 2.00 getting "locked up" in the windows platform at layer 7 or something like that per chatGPT. I tried to uninstall DUI but it failed to fully close or exit and finally I had to go into services and manually stop the service and then after a reboot (with DUI now uninstalled) the internet issues cleared up. I think I have seen / experienced something similar with previous DUI versions so might be unique to my system?
DefenderUI Free or Pro?
 
@danb Charge $70-80 for CyberLock and obtain a good chunk of API tokens for each user. Then make the feature to allow them to scan their entire hard drive. Reasoning: It is unlikely that a user will have multiple infections/hackers hiding in their machines, so it will boil down to a handful of malware/hackware. How many tokens will it take to analyze 5 pieces of malware/hackware? Then thats what you should allow for and charge for.

There are plenty of us who buy the ESETs, AppGuards, Deep Instincts of security apps. We buy quality.

I am all for helping out the needy and the hacked, just do it some other way, maybe a separate opensource venture.
 
Last edited:
@danb DUI free since I'm running Cyberlock (& I also run SiriusLLM with Cyberlock w/SiriusGPT integrated) -- iirc you said that was fine not too long ago) thanks, thanks, thanks!!! :D
Thank you for letting me know! There is a 0% chance that DefenderUI Free caused an issue. There were only minor changes to DefenderUI Free, since it does not have Sirius, and I reviewed all of the changes, and there is not a chance that one of the changes caused any issues.

The only thing I can think of is what I mentioned above... maybe the first release of the CyberLock 9.0 beta created some unwanted firewall rules. Other than that, I am certain something else unrelated to our software is causing this issue.
 
Thank you for letting me know! There is a 0% chance that DefenderUI Free caused an issue. There were only minor changes to DefenderUI Free, since it does not have Sirius, and I reviewed all of the changes, and there is not a chance that one of the changes caused any issues.

The only thing I can think of is what I mentioned above... maybe the first release of the CyberLock 9.0 beta created some unwanted firewall rules. Other than that, I am certain something else unrelated to our software is causing this issue.
ok thanks, the only reason I thought it was DUI, when I tried to uninstall it exiting it per instructions failed, I had to go into services to manually kill it. So I don't think it was DUI per se, but that it got "cross-tangled" in windows security stack. ChatGPT explained it in 50 words or less as: >> DefenderUI likely set Microsoft Defender ASR rules and related protections at the platform (policy) level rather than the user/UI level. Those settings persist in Group Policy/registry after uninstall, leaving Defender locked in a hardened state (e.g., ASR Level 7) that standard UI changes can’t override, causing browser/network functionality issues until policies are reset. << But yes could have been firewall rules you mentioned in initial Cyberlock 9.00 beta. (and chatgpt could be right or clueless...) Eager for Cyberlock 9.01 :)
 
  • Like
Reactions: danb and Dave Russo
False positive I believe:

Official website: https://www.drivermagician.com/DriverMagician.exe

VirusTotal diagnosis: VirusTotal

SiriusGPT 1.00 Beta diagnosis:

Total tokens: 0 (0 request / 0 response)

File path: d:\test\desktop\drivermagician.exe
File hash: 101bf19cd6b1a9b8019c06cd3ea467bbaf8dd2578613e8132eb54f84a946f99c
File size: 5.92 MB
File publisher: This file is a signable file type but has not been digitally signed.

Final Verdict: Not Safe with 92% confidence.

### Analysis Summary
The file presents itself as “Driver Magician Setup” (version info and filename). While the internal version block looks professionally filled and is linguistically neutral, the binary itself is completely **unsigned**—unusual for a commercial installer—and carries a **5.4 MB overlay** whose entropy is the maximum 8.0, indicating packed/encrypted/grafted data. The Import Address Table is modest and contains only orthodox Win32 APIs; nevertheless, the overlay, the lack of signature, and the entropy anomalies outweigh the apparent legitimacy of the metadata. Taken together, these traits are far more consistent with a wrapped malicious dropper (or PUP installer) than with a bona-fide vendor setup program.

### Detailed Analysis
- **No digital signature** on a file that purports to be commercial setup software is a red flag; reputable vendors invariably sign installers.
- **Overlay**: 5.4 MB (88 % of the file) with entropy = 8.0 – textbook indication of appended encrypted, compressed, or foreign PE data. Malware droppers routinely graft their real payload into the overlay and unpack it at runtime.
- **Import table** is small (142 APIs) and limited to standard user-mode libraries. There are no suspicious imports such as `VirtualAllocEx`, `WriteProcessMemory`, `SetWindowsHookEx`, or socket APIs; however, the payload hidden in the overlay could easily resolve those dynamically.
- **Exports**: Three symbols, two of which (`dbkFCallWrapperAddr`, `__dbk_fcall_wrapper`) are emitted by certain Delphi/Embarcadero debug helpers; the third (`TMethodImplementationIntercept`) is VCL-style RTTI. The presence of these exports is consistent with a Delphi-built stub whose real purpose is to unpack and launch the overlay.
- **Strings**: The printable strings are overwhelmingly Delphi/C++ Builder RTTI tokens (`TInterfacedObject`, `TMonitor`, `TTypeTable`, etc.). There are no URLs, IP addresses, or recognizable malware commands—again typical of an obfuscated stub.
- **Section layout**: Ten sections, mix of RX and RW flags, no obviously malicious section names, but section virtual/physical size mismatches plus the huge overlay reinforce the suspicion that something non-standard is stored after the legitimate PE image.
- **System defences**: ASLR and DEP are enabled, but that is now default for most toolchains and does not negate the other anomalies.

### Portable Executable Imports
All 142 imported routines come from `kernel32.dll`, `user32.dll`, `oleaut32.dll`, `version.dll`, `advapi32.dll`, and `netapi32.dll`. They cover process/thread management, file I/O, simple registry access (`RegOpenKeyExW`, `RegQueryValueExW`), and the usual Unicode/Ansi conversion routines. Nothing overtly stealthy is imported, but the overlay can supply any additional capability at runtime.

### Portable Executable Exports
Only the three Delphi-style symbols mentioned above. Their presence supports the thesis that this is a lightly modified Delphi launcher rather than a conventional installer.

### Portable Executable Strings
The top meaningful strings are all Delphi RTL/RTTI identifiers; no filenames, URLs, or C2 artifacts appear in the static string set. This scarcity is consistent with code that was compiled with full RTTI but whose real logic is encrypted or compressed in the overlay.

### Likely Type / Purpose
The binary is best characterised as an **obfuscated dropper** masquerading as a driver-update setup program. The huge, high-entropy overlay most likely contains the real payload (further malware or PUP bundle) that is extracted and executed after launch.

Malware type: Dropper
Malware name: Dropper.DriverMagician
Final verdict: Malicious with 92% confidence.
 
Last edited:
ok thanks, the only reason I thought it was DUI, when I tried to uninstall it exiting it per instructions failed, I had to go into services to manually kill it. So I don't think it was DUI per se, but that it got "cross-tangled" in windows security stack. ChatGPT explained it in 50 words or less as: >> DefenderUI likely set Microsoft Defender ASR rules and related protections at the platform (policy) level rather than the user/UI level. Those settings persist in Group Policy/registry after uninstall, leaving Defender locked in a hardened state (e.g., ASR Level 7) that standard UI changes can’t override, causing browser/network functionality issues until policies are reset. << But yes could have been firewall rules you mentioned in initial Cyberlock 9.00 beta. (and chatgpt could be right or clueless...) Eager for Cyberlock 9.01 :)
That is odd... DefenderUI warns the user to right click on the tray icon and exit the app before uninstalling. Oh well, I am not going to over think it ;). Please let me know how well the new versions do.
False positive I believe:

Official website: https://www.drivermagician.com/DriverMagician.exe

VirusTotal diagnosis: VirusTotal

SiriusGPT 1.00 Beta diagnosis:
Interesting, thank you for letting me know! I had never tried Driver Magician so I downloaded it and ran it via CyberLock Windows Sandbox and noticed that Smart Screen blocked it as well. That is one of the cool things about Sirius... it knows when a file is prevalent, and that always helps with the verdict, especially if it is signed with a verified sig. Sirius is also instructed to be conservative, especially when files are not signed. Thanks again!
 
Hey guys, here are the latest versions. I was going to be finished last night, then a user found a bug in the right click Windows Explorer context menu that was in pretty much all of the products, so that took a little while to fix.

These should be perfectly stable and there should not be too many more things we need to fix, but if you find anything, please let me know!

Hopefully everyone is doing great with the tokens by now, if not please give it a couple more days and if we think we need to increase the daily allocation we certainly will.

CyberLock 9.01
SHA-256: 2189e67ae90254310da14b3de4066c2317aa481ff6f64539528901ce98425ac0

SiriusGPT 1.01
SHA-256: 659307a4ed7122943eb0365affab07f90d397ecfce3efbde90a33f0f6ae0728e

SiriusLLM 1.01
SHA-256: ccb4992672bf8c5e4cdffe124f60902797c818f15053bb268a96953d64333383

DefenderUI 2.01
SHA-256: 8afc76658eb3f2f51248ffc9b15c427bd763c033af832bcb10ff012f00fdfded

DefenderUIPro 2.01
SHA-256: 5cf8dbadd1915ebc52e25f36ad75454a2774e8c08364f68f2b7a3e57cc4dd86f

DefenderUISilent 2.01
SHA-256: 172d94fd11e7ac540f61f877378242e0ae4a59c4605347cf64992189d91c64da

WDAC Lockdown 2.01
SHA-256: 6fe33bc0757854f2701072b08b6eca020445eefbe3547c6249ef99280a818529

Thank you guys!

Dan
 
That is odd... DefenderUI warns the user to right click on the tray icon and exit the app before uninstalling. Oh well, I am not going to over think it ;). Please let me know how well the new versions do.
@danb Yes for clarification DUI did exactly that -- "right click on the tray icon and exit the app before uninstalling" but DUI was "jammed" and would not close would not exit or close, which is why I had to go into services as admin and "stop" its service (& disable automatic) to kill it, then uninstalled it, reboot, and my internet returned. Some odd combo of events or apps somehow locked it up.
EDIT or you mentioned you "found a bug in the right click Windows Explorer context menu that was in pretty much all of the products, so that took a little while to fix" :unsure:
 
Last edited:
  • Like
Reactions: danb
"Please let me know how well the new versions do." @danb > installed Cyberlock 9.01 seems aok as is the internet. :D In settings | basic the Smart Firewall is set to Recommended and | SiriusGPT creates both inbound & outbound rules for not safe items. Also running SiriusLLM 1.01 portable -- all good. Holding off on DefenderUI 2.01 for now.
 
  • Like
Reactions: danb
Hey guys, here is a quick update. I just finished updating the API LLM models, so we are using a 2 new models, and 1 model we used in our previous model lineup, I also updated the prompt instructions. So far the results look amazing, but we will not know for another 2-3 days when we can review more results. BTW, some of you might have experienced a couple of odd / incorrect verdicts the last 2-3 days... that is fixed now ;). I was just trying different models and had around 20-30 incorrect verdicts because the new models I was testing did not work well for our use case.

The next versions of all of our software will be ready in the next 3-4 days, and at that point there will not be that much more to do for a while, except for tweaking the rules with the new Sirius implementation. So I might be able to take it easy soon ;). Thank you guys!

Edit: BTW, if you guys have any files where the verdicts were incorrect, please post or email me the SHA-256 hash and I will delete the result from the database so we can test again to see if the new model lineup and prompt instructions fixed the verdict. In other words, if you try to test a file with an incorrect verdict, it will return the previous result unless I delete the previous result from the database. Over time we will delete old results, but we do not want to delete them all at once because then too many tokens are spent all at once.
 
Last edited:
@danb
SiriusLLM detects a dropper in the Fort Firewall program. In my opinion, this is a false positive. Could you please comment on this detection?
 
@danb
SiriusLLM detects a dropper in the Fort Firewall program. In my opinion, this is a false positive. Could you please comment on this detection?
Very cool, thank you... this is the EXACT type of files I was asking for above. Basically I am wanting to test questionable / head scratching files that used to be false positives under the old Sirius, with the new Sirius that has a new model lineup and prompt instructions. And this is a perfect example!

The old Sirius did have a "false positive" for this file... I put false positive in quotes because Sirius is especially hard on files create by cybersecurity developers who do not sign their files. Sirius basically has the stance "Cybersecurity developers, of all developers, know how important it is to sign your files, and there is absolutely no excuse for cybersecurity developers to not sign their files. This file is NOT SAFE.".

So I was super curious to test with the new Sirius just to see, knowing that the verdict could EASILY go either way. The new Sirius returned this verdict: Final verdict: Safe with 70% confidence. So I was pretty happy, and then I realized that Sirius reduced its confidence to 70%, so at that point I was even happier.

If you guys have other files like this, please post them... these are super interesting and helps me tweak the models and prompt instructions.

Having said that... the results of the new Sirius continue to be truly astonishing. Sirius 1.0 was really good, but the new Sirius is on a totally different level. Remember how I was saying that to make Sirius better, all we have to do is update the model lineup and prompt instructions? Well, I knew it was going to get better, but the results so far have been almost perfect... although the old Sirius had great results as well. The difference is... with the new Sirius, it is much more decisive, so not only is it faster, but it uses less tokens.

Having said all that... unsigned files can go either way, especially from cybersecurity developers. Which is exactly the way it should be. If a file is not signed, we have no idea about when the file was created, who created it, if it was tampered with, and so on. It is not unlike finding a USB flash drive on the street and inserting it into your computer and executing all of the files. Thankfully most cybersecurity products are signed.

FYI, here is the old Sirius summary report:

Total tokens: 0 (0 request / 0 response)

File path: C:\Users\User\User\FortFirewall-3.19.9-windows10-x86_64.exe
File hash: c7a60c820f4e2509393607b34176605c43dd1d981684a2b6ab82336cc694877e
File size: 6.31 MB
File publisher: This file is a signable file type but has not been digitally signed.
WhitelistCloud verdict: Not Safe

Final Verdict: Not Safe with 85% confidence.

Analysis Summary
The file presents a set of mixed but ultimately suspicious indicators:
- No digital signature on a 6.5 MB setup installer dated 2018.
- Overlay entropy 8.0 over 6 MB of appended data—classic for InnoSetup bundles that can hide additional payloads.
- Import set is purely high-level Win32/CRT; no cryptographic or networking APIs are imported by the loader stub, yet the bundle could drop executables that do.
- Strings reveal genuine Inno Setup 6.x installer directives and Pascal error classes—consistent with a legitimate NSIS/Inno setup—but also an unpopulated OriginalFilename field and no verifiable publisher identity.
- WhitelistCloud flags it as malicious (albeit with a cautious tendency toward false positives).

Balanced against the absence of overt hostile imports/exports, the decisive points are:
1. Unsigned installer whose stated product name (“Fort Firewall”) is not matched by any reputable cert.
2. Six-megabyte high-entropy overlay ripe for stowing undetected payloads.
3. External reputation tip-off.

While the visible loader code is innocuous, the package as a whole is high-risk.

Detailed Analysis
1. File Authenticity
– Version info claims “Fort Firewall Setup” by Nodir Temirkhodjaev, but no signature means no proof of origin.
– OriginalFilename is blank—often omitted by packed or repackaged installers.

2. Entropy & Structure
– OverlaySize 6337912 bytes with entropy 8.0 → compressed/encrypted blob typical of Inno Setup “setup.exe” containers. Inno itself is benign, but malware frequently borrows it to smuggle secondary binaries.
– SectionEntropy values (6.38 in .text, 5.78 in .rdata) are normal; .data (.idata) entropy 2.30 shows unpacked stubs.

3. Imports
– 122 functions, all standard user-mode APIs; notable absence of WinInet, WinHttp, Crypt* or socket APIs.
– Presence of AdjustTokenPrivileges, LookupPrivilegeValueW and SeShutdownPrivilege strings suggests installer may restart system or elevate.

4. Strings
– Top helpful snippets:
– /SUPPRESSMSGBOXES /NOCANCEL /VERYSILENT → typical silent-install flags.
– Wow64DisableWow64FsRedirection → 64-bit-aware dropper behavior.
– Compressed block is corrupted → LZMA/decompression usage inside Inno.
– No hard-coded IPs, URLs, or known malware campaign markers visible in plain text.

5. Security Features
– ASLR & DEP enabled, no certificate table → good protections on stub, but unsigned payload remains dangerous.

Likely Type
Unsigned but Inno-based installer; whether it drops legitimate “Fort Firewall” or an abused copy is unverifiable. The risk resides in the unsigned, opaque bundle rather than the visible stub code.

Portable Executable Imports
All imports belong to kernel32/user32/advapi32/shell32/oleaut32 runtime. Key observations:
- Process/thread and file APIs (CreateProcessW, CreateFileW, VirtualAlloc, VirtualProtect) support unpacking/loading secondary binaries.
- Registry calls limited to RegOpenKey/RegQueryValue—no writes, consistent with simple setup queries.
- No cryptographic hashes or random-seed APIs imported by the loader stub. Malicious encryption would have to be inside the compressed Inno payload, not the stub.

Because the stub imports only generic APIs, its presence alone is inconclusive; risk is judged from external reputation and unsigned overlay.

Portable Executable Exports
No exports present—expected for an Inno Setup loader.

Portable Executable Strings (Top 15 relevant)
1. `/VERYSILENT`
2. `/SUPPRESSMSGBOXES`
3. `/NOCANCEL`
4. `/CLOSEAPPLICATIONS /RESTARTAPPLICATIONS`
5. `Wow64DisableWow64FsRedirection` / `Wow64RevertWow64FsRedirection`
6. `SeShutdownPrivilege`
7. `Compressed block is corrupted`
8. `USERPROFILE`
9. `GetUserDefaultUILanguage`
10. `InnoSetupLdrWindow`
11. `setupapi.dll`
12. `uxtheme.dll`
13. `dwmapi.dll`
14. `oleacc.dll`
15. `EOutOfMemory / EAccessViolation` (Pascal exceptions)

These reinforce legitimate Inno Setup behavior yet also show readiness for silent elevation and process restart—traits desirable for malware droppers.

Speculative Assessment
Likely an Inno Setup 6 bundle masquerading as “Fort Firewall” but lacking any publisher attestation. Until the compressed payload is extracted and inspected, treat the installer as a potential dropper.

Malware type: Dropper
Malware name: InnoDropper.FortFirewall
Final verdict: Malicious with 85% confidence.


And here is the new one...

Total tokens: 4364 (3673 request / 691 response)

File path: c:\users\user\desktop\fortfirewall-3.19.9-windows10-x86_64.exe
File hash: c7a60c820f4e2509393607b34176605c43dd1d981684a2b6ab82336cc694877e
File size: 6.31 MB
File publisher: This file is a signable file type but has not been digitally signed.

Final Verdict: Safe with 70% confidence.

## Analysis Summary
The provided executable file, fortfirewall-3.19.9-windows10-x86_64.exe, exhibits characteristics that suggest it is a legitimate installer for a firewall application. The file's imports, exports, and strings indicate functionality related to system configuration, file operations, and user interface interactions, which are typical for a setup program. However, the absence of a digital signature and some unusual strings raise concerns about the file's authenticity and potential security risks.

## Detailed Analysis
The file's imports include functions related to system information, file operations, and user interface interactions, such as `GetSystemInfo`, `WriteFile`, `MessageBoxW`, and `CreateWindowExW`. These imports are typical for a setup program and do not immediately suggest malicious activity. The presence of `RegQueryValueExW`, `RegOpenKeyExW`, and `RegCloseKey` imports indicates that the file interacts with the Windows registry, which is a common practice for installers.

The file has no exports, which is not unusual for an installer. The strings section contains a mix of expected and unexpected entries. The presence of `InnoSetupLdrWindow` and `STATIC` suggests that the file is built using the Inno Setup installer framework, which is a legitimate tool for creating Windows installers. However, some strings, such as `An unexpected memory leak has occurred` and `Compressed block is corrupted`, may indicate potential issues with the file's integrity or stability.

The digital signature section reveals that the file is not digitally signed, which is a concern. While the absence of a digital signature does not necessarily indicate malicious activity, it does make it more difficult to verify the file's authenticity and trustworthiness.

## Speculative Assessment
Based on the available attributes and features, it appears that the file is a legitimate installer for a firewall application. The presence of Inno Setup-related strings and the file's overall structure suggest that it is a setup program. However, the absence of a digital signature and some unusual strings raise concerns about the file's authenticity and potential security risks.

## Final Verdict
Given the analysis, the file's behavior and characteristics do not strongly indicate malicious activity. However, the absence of a digital signature and some unusual strings warrant caution. Since the confidence in the file being safe is higher than it being malicious based on the provided information, the verdict leans towards safety, but with a moderate confidence level due to the mentioned concerns.

Malware type:
Malware name:
Final verdict: Safe with 70% confidence.
 
One more thought, and this is kind of important, so please keep this in mind...

Other engines, like all of the VirusTotal engines, utilize signatures, heuristics, behavior and legacy Ai to answer this question... is this file potentially malicious?

Sirius utilizes LLM's and answers a very different question... is this file Safe to execute on my computer?

This is why the best use case for our products is serving as a pre‑execution filter for traditional antivirus or EDR solutions.
 

You may also like...