edited about scoring (#4)

Update inference and environment configuration

Refactor score handling with updated SCORE_EPSILON

Adjust SCORE_EPSILON and add reward safety clamp

Adjust SCORE_EPSILON and modify reward handling

Adjust SCORE_EPSILON and add reward safety clamp

Enforce strict open-interval task scores

Fix score bounds to strict open interval

Fix score range validation and update deployment config

Fix app entrypoint for OpenEnv validator

MedCodeRL v1.0 - Full OpenEnv spec compliance, 90 cases, shaped rewards, structured logging

added all files