Signature Clipping

"Signature clipping" isn't a widely-used term of art in Bitcoin. It's a catchall for ECDSA signature malleability: ways to take a valid signature and produce a different valid signature for the same transaction, which changes the txid without changing the meaning.

There were two practical vectors:

• Loose encoding. ECDSA signatures are wrapped in DER, which allows exactly one encoding of each signature. Bitcoin's original code checked signatures with OpenSSL, though, and OpenSSL also accepted encodings that strayed from DER, so the same signature could be written several ways. BIP 66, enforced from July 2015, made strict DER a consensus rule, so a block containing any other encoding is invalid.
• Inherent S-malleability. For any valid (r, s) ECDSA signature, (r, n-s) is also valid for the same message. The low-S rule requires s to be at most half the curve order, picking one form. BIP 62 proposed it as a consensus rule and was never deployed. Bitcoin Core has enforced it only as relay policy since version 0.11.1 (October 2015), and BIP 146, a later proposal to make it consensus, was closed. SegWit (BIP 141) has no low-S rule.

SegWit (BIP 141) closed the issue at the protocol level by moving signature data out of the txid calculation entirely. For SegWit and Taproot inputs, signature tweaks don't change the txid no matter what an attacker does.

For transactions that spend only SegWit or Taproot outputs, third-party malleability is solved. Transactions that spend older outputs, such as legacy P2PKH, can still be altered in principle. Strict DER is a consensus rule, while low-S is only relay policy. The network won't pass along a tweaked copy, but a miner could still put a high-S copy in a block.