ACCORDING to SANS[.]edu, IPv4-mapped IPv6 addresses are defined as a transition mechanism used to retain backward compatibility as IPv6 is deployed, and they are not used on the network but are translated to IPv4 before a packet is sent. To map an IPv4 address into IPv6, the prefix ::ffff:/96 is used, leaving the last 32 bits to represent the IPv4 address, for example turning 10.5.2[.]1 into ::ffff:0a05:0201, with many applications displaying the last four bytes in decimal for readability.
The article notes that whether these addresses can be used depends on the application, and it provides anecdotes from ping6 on macOS and Linux, ssh, curl, and browsers, illustrating that behaviour varies: some tools attempt IPv4 traffic, others may hang or fail with errors, while SSH and browsers can send traffic that appears as IPv4. It also points out that browsers may evade some filters by using IPv4-mapped IPv6 addresses when simple string matching is used, emphasising the need for careful filtering.
The discussion concludes that, on the network, the traffic ends up as normal IPv4 traffic, not IPv6 traffic using IPv4-mapped IPv6 addresses.