chore: removing logs

fix: remove exception

feat: rewards upgrade

fix: implementing strict prompt conditions for scores/reward to be in 0.0–1.0 range

feat: updating prompt and reward consdition

chore: updating the [END] log

feat: LLM agent model change

chore: update doc string

fix: clamp all rewards and scores to [0.10, 0.90]

fix: clamp all score paths to (0.01, 0.99), fix reward field name, add per-task score line

fix: harden grader prompts to prevent out-of-range scores

fix: reduced the number of scenario in inference

fix: logs in inference

chore: doc string update and remove unused import

fic: score condition

fix: score format and range

fix: score range

fix: enforce reward bounds (0.01–0.99) and 2 decimal precision across grader, env, and inference

fix: reward scores are updated to be between 0 and 1

chore: code cleanup

chore: logs format update

chore: updating logs

fix: rewards field shows single composite score instead of step CSV

chore: 2 sceanrios per difficulty - inference

chore: only one scenario is run per difficulty

chore: restrict stdout to START/STEP/END for eval compliance

fix: cast fallback action_type to Literal for Pylance compliance and remove image from repo root

docs: updating readme with state changes and test

feat: implement WhyDidItFailState for full OpenEnv state compliance

docs: Update README.md

docs: updating the readme with AI usage disclosure section

docs: adding detailed docs for agent_prompt , grade and scenarios

docs: expand setup section, fix factual errors, add features summary

feat: updating the readme

fix: minor refactor of env class name

chore: clean up all the unnecessary comments

Merge pull request #2 from samrat-rm/feat/llm-judge

chore: updating logs

feat: openEnv playground UI basic implementation

feat: updated the yaml file with tasks for evaluation

fix: comply with openenv stdout spec, preserve inspection data in history, sharpen medium-tier label rules

fix: error handling for episode run loop

fix: updating the logs to align with the evaluation standards

fix: error handling for HF_TOKEN env

feat: update the readme.md

fix: tighten label rules for underfitting, overfitting, and vanishing gradients

fix: normalize underfitting gradient norms and guard vague-answer penalty

feat: add judge fallback

refactor: moving llm judge inside server dir

feat: add playground static file serving