Put chosen tool bundles into tool/ directory.
python setup_eval.py with sample_test level dir.
bash run_infer_patches.sh to generate swe agent patches.
bash run_eval_patches.sh to evaluate patches. 