fix: make inference.py crash-proof with multi-strategy env connection

refactor: full openenv protocol compliance

refactor: move all root-level files to repo root

fix: ensure LLM calls use hackathon LiteLLM proxy env vars

fix: read API_KEY env var for LiteLLM proxy compatibility

fix: make POST /reset body optional for OpenEnv validator compatibility

Merge fix/codebase-hardening: packaging, grading, registry, tests, deploy

fix: packaging, grading baseline, action registry, observations, tests, deploy config

chore: untrack docs/ and scratchpad.md, add to .gitignore

Merge feat/real-world-utility-improvements: cross-column, profiling, undo, adaptive baseline, LLM prompt

feat: Unit 7 verified - LLM integration works with all 16 actions

feat: Unit 6 - comprehensive LLM prompt with all 16 actions, cost model, grading formula

feat: Unit 5 - adaptive baseline with positive scores on all 4 tasks

feat: Phase 1 - cross-column validation, observation enrichment, undo, profile

plan: real-world utility improvements -- cross-column, profiling, undo, adaptive baseline

fix(ci): use openenv CLI directly, not python -m

ci: add Docker build, unit tests, and inference smoke test workflow

fix: pass openenv validate - add uv.lock, server entry point, openenv-core dep

Merge feat/hackathon-compliance-enhancements: close hackathon compliance gaps + deep enhancements

fix: match reference inference.py - score 3dp, robust START/END pairing

chore: mark plan as completed

fix: smart fill_missing strategy - mode for categorical, median for numeric

feat: Units 14+15 - full inference with 14 tools, comprehensive README

feat: Unit 13 - expert task integration with ML metadata

feat: Units 11+12 - configurable difficulty, procedural gen, ML metadata

feat: Units 9+10 - cost-aware rewards, richer noise taxonomy

feat: Units 7+8 - sample rows, action history, 7 new action types

feat: Phase 1 + Unit 6 - deterministic grader, inference.py, Dockerfile, compliance fixes

feat: Data Cleaning RL Environment for OpenEnv Hackathon