unknown command -- whoami and dir

[CleverNamesTaken]

Hey! This looks like I really cool project, so thanks for creating it!

I'm new to Adaptix, as well as the API. I'm not sure what I need to do to resolve the issue that all of the commands appear to be unknown.

If I run whoami, or dir in the client GUI, it works fine.

Any idea what I might be missing?

Logging in to https://127.0.0.1:4321/endpoint as kali ...
✓  Logged in
No agent.id in config — using first available: 08a980d0
Agent  08a980d0  COMMANDO\kali  Windows  agent.x64.exe:6464  ⚡ elevated
[1/5] whoami
  ✗ DISPATCH FAILED  unknown command: command 'whoami' not found for agent 'beacon' os=1

[2/5] dir C:\Users\kali\test
  ✗ DISPATCH FAILED  unknown command: command 'dir' not found for agent 'beacon' os=1

[3/5] dir C:\Users\kali\test
  ✗ DISPATCH FAILED  unknown command: command 'dir' not found for agent 'beacon' os=1

[4/5] xyzzy frobnicate
  ⚠ XFAIL (dispatch failed)  unknown command: command 'xyzzy' not found for agent 'beacon' os=1

[5/5] bof /tmp/nonexistent.o
  ⚠ XFAIL (dispatch failed)  unknown command: command 'bof' not found for agent 'beacon' os=1

───────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────
  5  run              
  0  passed           
  3  dispatch-failed  
  2  xfail            
╭─────────────────────────────────────────────────────────  DISPATCH FAILED  ─────────────────────────────────────────────────────────╮
│ whoami  unknown command: command 'whoami' not found for agent 'beacon' os=1                                                         │
│ dir C:\Users\kali\test  unknown command: command 'dir' not found for agent 'beacon' os=1                                            │
│ dir C:\Users\kali\test  unknown command: command 'dir' not found for agent 'beacon' os=1                                            │
│                                                                                                                                     │
╰─────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────╯

[TheGr3atJosh]

Hi,
thank you for the bug report. I'm happy to help.
Could you share a bit more about your setup so I can try to reproduce this?
I just tried with a fresh AdaptixC2 and Extension-Kit installation and it worked for me.
The only way I was able to get the same output was when the Extension-Kit wasn't installed but then the GUI obviously didn't work either.
This is definitely interesting.

[CleverNamesTaken]

Thanks for looking!

Uncommenting out the line for extension-kit.axs in my profile.yaml appears to have fixed it.

    - "../Extension-Kit/extension-kit.axs"

[TheGr3atJosh]

Ah, that makes sense. But why did it work in the GUI if Adaptix wasn't loading your Extension-Kit before?

[CleverNamesTaken]

I think it is the difference between server-side loaded scripts and local scripts. ps list and ls work as expected.

It is weird though, because when I change my username to connect to the same server, my loaded local scripts are still available to me.

I'm not sure what is going on exactly, but it seems like the fix is simple -- load the scripts on server startup.

I guess you aren't seeing the same behavior on your end?

[TheGr3atJosh]

My apologies I misunderstood your problem. I assumed you had the Extension-Kit registered as a server script. When I register the Extension-Kit as a local script I do get the same behavior. However that is to be expected it even says so on the README:

Note: Only commands supported by the server-side AxScript engine work here. Client-side hook commands are not available. Some commands complete successfully but return no text output via the task list API (e.g. fs pwd on Kharon) — omit expected for those.

You can't run local ax scripts via the Adaptix API. In fact I initially tried to write this Testing-Kit as an .ax script but that was super messy. The scripts only trigger when the AdaptixClient is actually open, there is no way to serialize/wait, debugging is very inconvenient, no LLM has any idea how they work, etc. In my experience you are almost always better off registering scripts you want to permanently use (like the Extension-Kit ) at the Server side (unless you just want to do a quick test).
If you want to do automation I would stick with Python + API (although that site is a bit outdated.)
Let me know if I can make that clearer in the README.

[CleverNamesTaken]

Ha, no, it is very clear in the documentation.

As is often the case, failure to read leads to avoidable problems.

Thanks so much!