Three Gates

; z4_multi_stage: "Three Gates". A staged check that chains three independent
; tests; you must satisfy all of them with one input, so reverse them in order:
;   Stage 1 (length): the input must be exactly 7 bytes.
;   Stage 2 (sum):    the byte values must add up to 505 (0x1F9).
;   Stage 3 (xor):    all seven bytes XOR-folded together must equal 0x53.
; Technique: multi-stage check. r0 = 1 on accept. Accept rule: nonzero.
;
; 8-BIT NOTE: the true byte sum 505 overflows a byte; the byte core keeps the
; running sum mod 256, so the sum gate is the LOW BYTE of 505 (505 & 0xFF = 0xF9).
; The technique (a length gate, then a sum gate, then an XOR gate, all immediates)
; is unchanged; only the sum's width is reduced to the 8-bit ceiling.
;
; sample "NEXUS42": N+E+X+U+S+'4'+'2' = 505 (low byte 0xF9); XOR-fold = 0x53.

        ; --- stage 1: length gate ------------------------------------------
        len   r2
        cmp   r2, 7
        jnz   bad

        ; --- stages 2 + 3: one pass building both the sum and the XOR -------
        mov   r3, 0         ; r3 = running byte sum (mod 256)
        mov   r4, 0         ; r4 = running XOR fold
        mov   r7, 0
sloop:  cmp   r7, 7
        jge   final
        ldb   r0, [r7]
        add   r3, r0        ; sum += byte
        xor   r4, r0        ; xor-fold ^= byte
        inc   r7
        jmp   sloop

final:  cmp   r3, 0xF9      ; stage 2: low byte of 505
        jnz   bad
        cmp   r4, 0x53      ; stage 3: XOR fold
        jnz   bad
        mov   r0, 1
        ret
bad:    mov   r0, 0
        ret