A double-strand break is the easy part; what you do with it decides the edit. US10526590B2 - "Compounds and methods for CRISPR/Cas-based genome editing by homologous recombination," issued January 7, 2020 to Agilent Technologies - claims the editing method anchored to the homology-directed repair (HDR) pathway, and that anchoring is the claim's center of gravity.
Cells repair a Cas-induced break by one of two dominant routes: error-prone non-homologous end joining (NHEJ), or precise homology-directed repair using a supplied template. A claim that recites editing 'by homologous recombination' carves out the precise-repair route specifically. That is narrower than a claim to any CRISPR cut, and it is deliberately so - HDR-based precise correction is the commercially valuable mode for fixing point mutations.
The CPC assignment is spare here - C12N 9/22 alone - which tells you the office read this primarily as a Cas-nuclease-class invention applied to a repair method, not as a sprawling composition family. A spare CPC profile often signals a tightly-scoped claim set rather than a defensive thicket.
Why does the repair-pathway limitation decide so much? Because a competitor practicing CRISPR editing that relies on NHEJ-based disruption - knocking a gene out rather than precisely correcting it - is arguably outside a claim drawn to homologous recombination. The limitation both narrows the claim past the dense cutting-machinery prior art and confines enforcement to programs doing templated correction.
For the landscape, this is a reminder that 'CRISPR patent' is never a sufficient description. The estate fractures by mechanism - cutting, base editing, prime editing, NHEJ disruption, HDR correction - and a grant tied to one repair pathway reaches only the programs that use it.