Add openswe_v3.yaml: Phase 4 128K context test (max_input_tokens=131072)

Upload deepswe_repro.yaml with huggingface_hub

Add openswe_v2.yaml: author-provided config (function_calling, drop diff_state, max_input_tokens=32768)

Upload qwen3_acc_thinking.jinja with huggingface_hub

Upload openswe.yaml with huggingface_hub

upload SERA paper sweagent e2e.yaml config

initial commit