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: 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: logs format update

feat: implement WhyDidItFailState for full OpenEnv state compliance

chore: clean up all the unnecessary comments

feat: openEnv playground UI basic implementation

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

fix: harden label rules to prevent missing_regularization misfires

feat(scenarios): add real gradient norms and improve scenario discriminability

feat: upgrading the inspect feedback function

feat: updating the logs with relevant model names for improving score function efficiency

feat: updating the evidence scoring function

feat: adding steps count logic to encourage the agent explore more

feat: max step limit

feat: fix suggestion is required and not providing fix causes penalty

feat: ordering_bonus function implementation

feat(grade): inspected is upgraded to inspected_order. It rewards steps taken in order

refactor: Aligning the env with the new grade function

feat: adding more scenarios

feat: grade function refactor and additional features

fix: add state() method

feat: init graders and implement grade_easy() in env

chore: import statement refactor

fix: seed and episode_id in reset()

fix(grade): keyword matching and requires_fix flag for diagnosis scoring

feat: initial environment setup

refactor: WhyDidItFailEnvironment class name

refactor: WhyDidItFailAction and WhyDidItFailObservation classes

feat: Initialised sceanrios

Initial commit