AbhishekKharat11/disaster_response_env

fix: clamp ALL scores strictly within (0.01, 0.99) in grader, environment, app, inference - Phase 2 Task Validation fix

fix: clamp all scores strictly within (0.001, 0.999) — Phase 2 Task Validation fix

fix: HF_TOKEN no default, add LOCAL_IMAGE_NAME env var (checklist compliance)

Fix inference.py: exact START/STEP/END format with OpenAI client

Fix HF port + uvicorn binding

Fix validator: script entry + uv.lock

Add main() entrypoint for validator

Add inference.py at repo root

Grader v3: allocation-priority alignment, hard contradiction penalties, state mutation

Upgrade grader: semantic scoring, contradiction detection, temporal awareness

Add root endpoint

Simplify baseline agent to use /baseline endpoint

Fix action wrapper format for step endpoint

Fix step action format

Rewrite baseline agent - no API key required, uses HTTP directly

Add openenv tag

Add openenv tag to Space

Add /tasks /grader /baseline endpoints - required for submission

Move Dockerfile to root for HF Spaces

Fix colorTo metadata

Add HF Space metadata

Initial commit - Disaster Response Environment