Optimize inference.py: never break loop early, switch to heuristic after max_steps to ensure completion within time limit

Optimize inference.py to prevent timeout: add 8s LLM timeout, cap hard task steps, and add global timer fallback

Fix openenv validate error: add openenv-core to dependencies

Fix 422 error by making body optional in reset and step

Merge branch 'main' of https://github.com/vettri06/meta_ai_hackathon

Fix critical submission issues: API_KEY, API_BASE_URL, and [END] format

Delete _agents directory

Delete REVIEW_AND_TODO.md

Delete implementation_plan.md

Delete progresss.md

Update default model to Qwen-7B

Fix 404 by adding root route

Deploy to Hugging Face Space with correct permissions

Final submission prep

Initial commit