Log coverage metrics show Claude Opus 4.6 tests uncover 28.4% more unique log templates than human-written Locust tests on Light-OAuth2, while combinations of strategies increase observed coverage by 30-105%.
Cogo, and Shane McIntosh
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.SE 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Assessing REST API Test Generation Strategies with Log Coverage
Log coverage metrics show Claude Opus 4.6 tests uncover 28.4% more unique log templates than human-written Locust tests on Light-OAuth2, while combinations of strategies increase observed coverage by 30-105%.