How to create an action?

This document helps you to create an action for events

Actions are an essential part of the events.

Each event consists of zero or multiple actions. An event with zero actions is treated as a disabled event.

There are three types of actions in HyperDbg as described here.

You can have multiple "Custom Codes", "Script", and "Break".

This document is a brief of how to create actions for an event.

Pre-Allocated Buffers

There are many limitations when your script is running in vmx-root. To address these limitations, HyperDbg employs pre-allocated buffers.

Pre-allocated buffers are a buffer that HyperDbg previously allocated from the non-paged pool resource of the system as it is dedicated to being used within the events.

You can use these buffers safely from vmx-root to save your temporary variables or needed data. Keep in mind that there is only one pre-allocated buffer for an event, so if you want to access it from different cores, you should consider using spinlock functions to avoid concurrency.

Break

Break to the debugger, works exactly like classic debuggers like Windbg.

If you simply use the command without any extra parameters, it will be treated like classic debuggers, and HyperDbg gives the system control to the debugger.

Script

Custom vmx-root mode compatible script engine is another feature for HyperDbg.

You can use scripts within events by specifying your script within script { and } in the event's command.

Note that when the event is triggered, you can modify memory and registers, and when each event is triggered, it has its own set of registers, context, and memory layout. You should keep in mind that each event might be triggered simultaneously within different cores.

A pointer to the pre-allocated buffer for the target event is available in the $buffer pseudo-register.

There are different examples of using the script engine effectively within events here.

Custom Codes

Run custom code lets you run your custom assembly codes whenever a special event is triggered; this option is fast and powerful as you can customize the HyperDbg based on your needs.

Accessing random memory in custom code and condition code is considered "unsafe". You have some limitations on accessing memory on some special events.

Run custom code without a safe buffer

Each command in HyperDbg that are tagged as "event" in the document follows the same structure described here. At the time you execute a command, you can add a code { xx xx xx xx } where xx is the assembly (hex) of what you want to be executed in the case of that event.

Generally, the assembly code in the code block will be called in the following form.

typedef PVOID
DebuggerRunCustomCodeFunc(PVOID PreAllocatedBufferAddress, PGUEST_REGS Regs, PVOID Context);

As it called in the fastcall calling convention, PreAllocatedBufferAddress will be on rcx, Regs will be on rdx and Context is on r8.

PreAllocatedBufferAddress is the address of a non-paged safe buffer, which is passed to the function on rcx. (more about it later).

Regs, for general-purpose registers, we pass a pointer to the following structure as the second argument on rdx.

typedef struct _GUEST_REGS
{
    ULONG64 rax; // 0x00
    ULONG64 rcx; // 0x08
    ULONG64 rdx; // 0x10
    ULONG64 rbx; // 0x18
    ULONG64 rsp; // 0x20 
    ULONG64 rbp; // 0x28
    ULONG64 rsi; // 0x30
    ULONG64 rdi; // 0x38
    ULONG64 r8;  // 0x40
    ULONG64 r9;  // 0x48
    ULONG64 r10; // 0x50
    ULONG64 r11; // 0x58
    ULONG64 r12; // 0x60
    ULONG64 r13; // 0x68
    ULONG64 r14; // 0x70
    ULONG64 r15; // 0x78
} GUEST_REGS, *PGUEST_REGS;

The Context is a special variable that shows an essential parameter of the event. This value is different for each event. You should check the documentation of that command for more information about the Context. For example, Context for !syscall command is the syscall-number or for the !epthook2 command is the physical address of where the hidden hook triggered. Context is passed to the custom code as the third argument on r8 .

PreAllocatedBufferAddress (rcx) is always NULL in Run custom code without a safe buffer, and it's used in Run custom code with a safe buffer.

As an example, we want to find the TAG (ExAllocatePoolWithTag). If the tag is a special value, then we want to change it to a new value.

As you know, ExAllocatePoolWithTag in Windows is defined as:

PVOID ExAllocatePoolWithTag(
  POOL_TYPE                                      PoolType,
  SIZE_T                                         NumberOfBytes,
  ULONG                                          Tag
);

Based on the x64 calling convention, the parameters are passed as rcx, rdx, r8, r9, and stack and Tag is on r8.

As you know, if you want to change a register in the target OS, you have to find the register in Regs and change it from there. Based on _GUEST_REGS, r8 is on 0x40 from the top of this structure.

Take a look at the following assembly code. It first checks whether the Tag (r8) is HDBG, and if it's HDBG, then we change it to HDB2.

When we convert the above code to assembly, then we have the following code :

0:  48 8b 5a 40             mov    rbx,QWORD PTR [rdx+0x40]
4:  48 81 fb 48 44 42 47    cmp    rbx,0x47424448
b:  74 02                   je     f <ChangeIt>
d:  eb 08                   jmp    17 <Return>
000000000000000f <ChangeIt>:
f:  48 c7 42 40 48 44 42    mov    QWORD PTR [rdx+0x40],0x32424448
16: 32
0000000000000017 <Return>:
17: c3                      ret

Imagine, the ExAllocatePoolWithTag is located at fffff800`4ed6f010. We can hook and change the Tag using the following command.

HyperDbg> !epthook2 fffff800`4ed6f010 code {488B5A404881FB484442477402EB0848C7424048444232C3}

Run custom code with a safe buffer

The difference between "Run custom code without a safe buffer" and "Run custom code without a safe buffer" is that you have an extra parameter, called buffer xx where xx is the hex length of the buffer.

The PreAllocatedBufferAddress is just one buffer. You have to know how many cores you have. If there are two or more cores that might use the buffer simultaneously, you have to use a special location (offset from the top of the buffer) for each core to avoid race conditions and unintended behavior.

You can use the buffer which is available in rcx.

Last updated