[Mempool spam] Should we as developers reject non-standard Taproot transactions from full nodes?

Weiji, a member of the bitcoin-dev mailing list, recently proposed a new opcode called OP_ZKP to enable payments based on zero-knowledge proof. He believes it could be a viable way to address transaction fee issues and enable more smart contracts on Bitcoin's native L2. Although he hasn't figured out how to use OP_ZKP to incentivize BRC-20, inscription etc. to move to L2, he asks for feedback and comments from the group. In response to Weiji's proposal, Erik Aronesty emphasized that he is not an advocate for doing anything at this point aside from working on L2. However, he finds the discussion of outputs and fees interesting and related to the game theory portion of Bitcoin. Jaroslaw points out that the spam itself is not the real reason for the recent congestion in Bitcoin's mempool, but rather the $30 fee per transaction. He quotes several participants in the discussion who argue that action against exorbitant fees should have been taken months ago and that a solution must be found that fits everyone's common ground. Luke Dashjr suggests extending existing filters or trying a more narrow approach like OP_RETURN to curb spam transactions, which threaten the smooth and normal use of Bitcoin. Ali Sherief argues that if the volume of spam does not die down over the next few weeks, similar action as during the block size wars and Segwit might need to be taken now. He proposes enforcing censorship at the node level and introducing a run-time option to instantly prune all non-standard Taproot transactions.