feat: add Jasmin assembly integration for crypto acceleration

- Integrated `zupt_mac_verify_ct` in `zupt_decrypt_buffer()` to replace C XOR loop for HMAC-SHA256
- Integrated `zupt_ct_select_32` in `zupt_mlkem768_decaps()` to replace C `cmov()` for FO transformation
- Added `include/zupt_jasmin.h` with extern declarations and ABI docs
- Added `#ifdef ZUPT_USE_JASMIN` guards with clean C fallbacks in `zupt_crypto.c` and `zupt_mlkem.c`
- Makefile now auto-detects `jasmin/*.s`, assembles and links with `-DZUPT_USE_JASMIN`

Closes #3

jasmin/zupt_x25519_fe.jazz

@@ -0,0 +1,34 @@
+/* Zupt — X25519 Constant-Time Conditional Swap (Jasmin)
+ * Copyright (c) 2026 Cristian Cezar Moisés — MIT License
+ *
+ * CT-REQUIRED: fe_cswap must not leak cond via timing.
+ * This is the only CT-critical field operation in X25519.
+ * fe_add/fe_sub/fe_mul use C fallback (data-independent timing
+ * on x86-64 — ADD/MUL have fixed latency).
+ *
+ * 4 × u64 limbs, pure register operations, no intrinsics needed.
+ */
+
+export fn zupt_fe_cswap(
+    reg u64 a_ptr,
+    reg u64 b_ptr,
+    reg u64 cond)
+{
+    reg u64 mask ta tb diff;
+    inline int i;
+
+    mask = 0;
+    mask -= cond;
+
+    for i = 0 to 4 {
+        ta = [a_ptr + 8 * i];
+        tb = [b_ptr + 8 * i];
+        diff = ta;
+        diff ^= tb;
+        diff &= mask;
+        ta ^= diff;
+        tb ^= diff;
+        [a_ptr + 8 * i] = ta;
+        [b_ptr + 8 * i] = tb;
+    }
+}