fix: clamp scores to (0.001, 0.999) for Phase 2 strict range validation

fix: add structured stdout tags and flush for Phase 2 grader

fix: explicit setuptools package discovery

feat: restructure repo for OpenEnv compliance

chore: add pyproject.toml for multi-mode deployment compliance

fix: make reset endpoint body optional for automated judge

fix: hardened JSON parsing and env var fallbacks for automated judging

feat(ui): implement autopilot visual streaming and dynamic state injection

feat: complete v2 React UI rebuild and Hugging Face integration

feat(core): finalize v2.0 architecture with JSON Healer, Nemotron-120B integration, and Command Center UI

feat: bulletproof /autopilot with server-side JSON sanitizer

feat: upgrade to Nemotron-120B with server-side LLM proxy

feat: Add secure UI input for LLM API key

feat: fuse React SRE Command Center UI with FastAPI backend

feat: switch to NVIDIA API catalog and fix LLM token limits

feat: complete contain and investigate environment engine

feat: refactor evaluation script to support remote API interaction and structured JSON responses

Add Hugging Face config to README

First OpenEnv deploy