Blazing ASM

<img src="https://github.com/user-attachments/assets/562c7058-7e2e-40ff-be93-7a3b57d940cb" width="100" />

A simple and lightweight assembler library that supports the majority of significant instructions from the Intel x86-64 architecture. Mainly beneficial for shellcode development. Due to its design, it can produce the machine code quite fast while allowing you to use random registers/immediates at runtime.

Requirements

• C++20 or above
• STL is not required (Compatible with kernel driver projects)

Why not use AsmJit?

If you need to generate x86-64 machine code with minimal overhead, BlazingASM is a powerful alternative to dynamic assemblers like AsmJit. Designed for compile-time code generation, it emits raw and fixed-size machine instructions directly into static arrays, eliminating the need for runtime encoding or memory management. Single C++ header makes it really easy to integrate ~ even in kernel or embedded system projects. With a predictable layout, it's able to construct instructions at compile-time while handling dynamic operands at runtime without overhead. Particularly useful in exploit development, low-level code mutation and obfuscation, penetration testing and any scenario where shellcode is involved.

That being said, BlazingASM is intentionally minimal and template focused. It doesn’t aim to match AsmJit’s vast instruction coverage, runtime flexibility, or feature set. Instead, it trades breadth for simplicity and performance ~ making it a practical tool when you know exactly what code you need, and want it fast and small.

There are two different approaches you can utilize:

Static Mode

For a better demonstration, below is a screenshot of the comparison between the code and the disassembly (with assemble_static()):

Dynamic Mode

If you wish to use the dynamic code generation, below is a screenshot of the comparison between the code and the decompilation on IDA (with assemble()):

Instructions Supported

Category	Instruction	Description
Data Movement	MOV	Register/Memory/Immediate moves
	LEA	Load effective address
	DB(...)	Define raw byte(s) (data injection)
Arithmetic (GRP1)	ADD, OR, ADC, SBB, AND, SUB, XOR, CMP	Basic arithmetic and bitwise operations
	TEST	Bitwise AND for flag setting (no result stored)
Shifts/Rotates (GRP2)	ROL, ROR, RCL, RCR, SHL, SAL, SHR, SAR	Support both constant and 1-bit variants (ROL1, ROR1, etc.)
Unary Ops (GRP3)	NOT, NEG	Logical negation, two's complement
	MUL, IMUL	Unsigned and signed multiplication
	DIV, IDIV	Unsigned and signed division
Control Flow	JCC	Conditional jumps via Conditions enum
	JMP, JMP32	Unconditional jump (near and 32-bit override)
	CALL, CALL32	Near calls (with 32-bit override support)
Stack Operations	PUSH, PUSH32, POP, POP32	Push/pop values (supports 64-bit and 32-bit override)
	POPF, PUSHF	Push/pop flags
Fixed Instructions	RET	Return from procedure
	NOP	No operation
Planned	FPU Ops, INC/DEC	🚧 Coming Soon...

Setting it up

All you need to do is download the blazing_asm.hpp header file, paste it into your project folder, and write this into your source code:

cppCopy
#include "blazing_asm.hpp"

.
.
.

auto shellcode = basm::assemble_static (
  basm::MOV (basm::RAX, basm::RBX),
  basm::MOV (basm::RCX, basm::RDX)
);

You can also use basm namespace within a scope for a simpler syntax:

cppCopy
{
  using namespace basm;
  
  auto shellcode = assemble_static (
    MOV (RAX, RBX),
    MOV (RCX, RDX)
  );
}

Usage

You can find the detailed explanation in the Documentation Page.

🚧 Note

This project is currently in development, so if you come across any problems, please create an issue. Feel free to contribute and improve.