Short answer: I don’t know of a well‑documented case in the…

Short answer: I don’t know of a well‑documented case in the last 10 years where Bitcoin mainnet produced a consensus‑invalid block that economic/full nodes rejected and that rejection caused a reorg of the then‑longest chain. The canonical examples people cite (2010 overflow bug, 2013 LevelDB split, 2015 BIP66) are older.