The New Jersey Machine-Code Toolkit helps programmers write applications that process machine code. Applications that use the toolkit are written at an assembly-language level of abstraction, but they recognize and emit binary. Guided by a short instruction-set specification, the toolkit generates all the bit-manipulating code. The toolkit's specification language uses four concepts: fields and tokens describe parts of instructions, patterns describe binary encodings of instructions or groups of instructions, and constructors map between the assembly-language and binary levels. These concepts are suitable for describing both CISC and RISC machines; we have written specifications for the MIPS R3000, SPARC, and Intel 486 instruction sets. Complete specifications for the three architectures appear in [ramsey:tk-architecture]. Excerpts from those specifications appear in this paper. The reference manual for the toolkit [ramsey:tk-reference] gives a complete description of the specification language. We have written two applications that use the toolkit: a retargetable debugger and a retargetable, optimizing linker. The toolkit generates efficient code; for example, the linker emits binary up to 10% faster than it emits assembly language, making it 1.5--2 times faster to produce an a.out directly than by using the assembler.