Thanks for writing this concise and clear document. The addition of TLV extensions seems useful.

I don't see any major issues, but there are a few minor ones.

Like the genart reviewer, I believe this "should" ought to be normative:

"When this extension mechanism is used, the number of Group Records in each Report message should be kept small enough that the entire message, including any extension TLVs can fit within the network MTU."

In Section 4, I suggest rewording this and fixing capitalization:

"To check this, one will need to walk through each of The TLVs until there are less than four octets left in the IP payload."

to:

"To check this, the parser needs to walk through each of the TLVs until there are less than four octets left in the IP payload."

In Section 5, I suggest rewording this:

"IGMP and MLD implementations, host implementations in particular, rarely change, and it is expected to take a long time for them to support this extension mechanism."

to:

"IGMP and MLD implementations (particularly implementations on hosts) rarely change, and the adoption process of this extension mechanism is expected to be slow."

I also suggest that you consider giving guidance for preventing this new extension point from ossifying, given the slow rate of change. Please see the discussion in RFC 9170. For example, if you cannot guarantee active use and change, you may want to suggest synthetic use of the extension point ("greasing" by endpoints sending unknown type values every now and then to make sure receivers correctly handle unknown types).