feat(swesmith): add SWE-Smith benchmark evaluation scaffold#428

Merged

juanmichelini merged 7 commits intoOpenHands:mainfrom

mkaramuk:feat/swe-smith

Apr 14, 2026

Commits on Apr 13, 2026

feat(swesmith): add SWE-Smith benchmark scaffold

reisepass
and
mkaramuk
committed
docs(swesmith): add prompt template, README, and env example

reisepass
and
mkaramuk
committed
fix(utils): support .json files in dataset loader

mkaramuk
and
reisepass
committed
fix: update swesmith dep to PyPI release and adapt to n_critic_runs rename
mkaramuk
committed
fix(swesmith): rename image_exists to remote_image_exists
mkaramuk
committed
test(swesmith): add test data and mocks for metrics collection test
mkaramuk
committed
fix(swesmith): use swesmith registry for mirror URL with public/private detection
mkaramuk
committed