Perfected the logs

openenv.yaml format fix & graders fix & inference.py fix output format

Vital changes from Soham

fix: finalize OpenEnv hackathon compliance and boundary requirements

Changed .yaml to link to reward functions and made reward functions in graders out of their classes

new fix

added the run_episode function

tweaked the minimum reward

moved code into all the grader flies from just one file and fixed the import name 'FinAuditorGrader' which was used for every task difficulty

fix: added the template repo structure of tasks/ and grader/