fix: broken uint256_t implicit copy (AztecProtocol#3625)

Look at the diff, then consider this code
```
#include <iostream>
#include <concepts>

class Uint256_t {
public:
    // Conversion operator (explicit)
    template <std::integral T>
    explicit operator T() const {
        std::cout << "Uint256_t conversion operator called\n";
        return static_cast<T>(data[0]);
    }
    int data[4] = {0,0,0,0};
};

class Field {
public:
    Field() {}
    // Constructor that takes a Uint256_t
    Field(int x) {
        std::cout << "This is silly!!!!" << std::endl;
    }
    Field(const Uint256_t& input) {
        (void)input;
        std::cout << "Field constructor called\n";
        // Additional initialization code here
    }
};

class Toot {
    public:
    Toot(Field f) {
          (void)f;
    }
};

int main() {
    Uint256_t uint256_value;
    Field x = uint256_value;
    Toot t{uint256_value};

    return 0;
}
```

With std::integral concept on the cast:

```
Field constructor called
Field constructor called
```


but without! which was caught in practice passing a uint256 to
FieldSponge

```
Field constructor called
Uint256_t conversion operator called
This is silly!!!!
```

This truncated the field mysteriously! Open-ended implicit conversions
are dangerous

---------

Co-authored-by: ludamad <adam@aztecprotocol.com>

barretenberg/cpp/src/barretenberg/numeric/uint128/uint128.hpp

@@ -5,6 +5,7 @@
 
 #ifdef __i386__
 #include "barretenberg/common/serialize.hpp"
+#include <concepts>
 
 namespace numeric {
 
@@ -37,7 +38,7 @@ class alignas(32) uint128_t {
     constexpr ~uint128_t() = default;
     explicit constexpr operator bool() const { return static_cast<bool>(data[0]); };
 
-    template <typename T> explicit constexpr operator T() const { return static_cast<T>(data[0]); };
+    template <std::integral T> explicit constexpr operator T() const { return static_cast<T>(data[0]); };
 
     [[nodiscard]] constexpr bool get_bit(uint64_t bit_index) const;
     [[nodiscard]] constexpr uint64_t get_msb() const;

barretenberg/cpp/src/barretenberg/numeric/uint256/uint256.hpp

@@ -14,6 +14,7 @@
 #include "../uint128/uint128.hpp"
 #include "barretenberg/common/serialize.hpp"
 #include "barretenberg/common/throw_or_abort.hpp"
+#include <concepts>
 #include <cstdint>
 #include <iomanip>
 #include <iostream>
@@ -91,7 +92,7 @@ class alignas(32) uint256_t {
 
     explicit constexpr operator bool() const { return static_cast<bool>(data[0]); };
 
-    template <typename T> explicit constexpr operator T() const { return static_cast<T>(data[0]); };
+    template <std::integral T> explicit constexpr operator T() const { return static_cast<T>(data[0]); };
 
     [[nodiscard]] constexpr bool get_bit(uint64_t bit_index) const;
     [[nodiscard]] constexpr uint64_t get_msb() const;

barretenberg/cpp/src/barretenberg/stdlib/primitives/bigfield/bigfield.fuzzer.hpp

@@ -266,7 +266,7 @@ template <typename Builder> class BigFieldBase {
                 mask = (uint256_t(1) << mask_size) - 1;
                 // Choose the bit range
                 // Return instruction
-                return { .id = instruction_opcode, .arguments.element = Element(temp & mask) };
+                return { .id = instruction_opcode, .arguments.element = Element(static_cast<uint64_t>(temp & mask)) };
 
                 break;
             case OPCODE::RANDOMSEED: