docs: refresh README for reviewers

final push

docs: updated blog.md

Add documentation for Sakha project

feat(grpo): enhance action response parsing by removing reasoning blocks and refining regex handling

feat(grpo): update max completion length and refine prompt handling for improved evaluation

feat(grpo): adjust training parameters and disable thinking mode for consistent action calls

feat(grpo): enhance training dynamics with new replay policies and update state steps

feat(grpo): update max sequence length and refine prompt formatting in training scripts

refactor: remove SakhaEnvWrapper class and streamline reward function in GRPO training script

refactor: simplify GRPO training script by removing CLI presets and syncing with notebook implementation

feat(notebook): add Unsloth 4-bit training and base-vs-trained LLM eval

feat(eval): consolidate eval_harness into eval_policies, add LLM policy support

refactor(eval): extract shared eval constants and policies into eval_common.py

docs(readme): fix action names, add Results section, update baseline scores

docs(readme): fix action names, add Results section, update baseline scores

feat(plots): add plot generation script and training evidence plots

fix(notebook): update Colab for PyPI install, full-shift training, baseline eval

feat(train): add CLI args (learning-rate, batch-size) and full-shift defaults

data(baseline): add reproducible baseline artifacts for easy/medium/hard

feat(eval): add reproducible eval harness with comprehensive metrics

test(rubric): add golden parity tests for rubric migration validation

data(fixtures): capture pre-migration golden reward fixtures for parity testing

feat(rubric): integrate SakhaRubric into SakhaEnvironment step/reward path

feat(rubric): add composable rubric scaffolding wrapping existing reward logic

fix(deps): migrate openenv-core from git URL to PyPI >=0.2.3

chore: pre-merge cleanup for colab-training branch

fix: remove stray brace causing invalid JSON in notebook

fix: add docstrings to ALL methods for TRL JSON schema parsing

fix: use proper multi-line docstrings for TRL JSON schema parsing

fix: install transformers and trl together to prevent version downgrade

feat: add GRPO training script with mode presets, eval split, checkpointing

scripts: add GRPO training demo for Sakha env

docs: add hackathon round2 checklist

fix(inference): default episodes to 1 for dev speed

docs: rewrite AGENTS.md with agent quick guide

fix(inference): align checklist compliance and structured run logs

chore: add artifacts/ to gitignore

test(inference): add prompt profile and rank_candidates tests

fix(eval_policies): ensure output directory exists before writing JSON

docs(hackathon): add Phase 2 stdout parsing requirements

feat(inference): integrate formatters, prompt profiles, and rewrite deterministic policy

feat(formatters): add structured output formatting system

docs: update README for new workflow

feat(tests): update for new action system

chore(server): update imports

feat(inference): adapt to new action system

feat(graders): implement new scoring logic

feat(env): implement workflow-based simulation

feat(models): added new action types and pending task system