chore: Apply Bug #2 and Bug #3 strict min/max bound clamping to prevent out of range scores and fix windows encoding

Sync Web UI [END] logs to use the exact formal score= specification format

Fix Phase 2 OpenEnv validation traps: add grader paths to openenv.yaml and safe parameterless defaults

Spec-compliance overhaul: remove difficulty_multiplier, weighted blend scoring, dep_hard fix, [END] format

Fix dep_hard Counter bug, add fatal error handling, update README with 14-model benchmark

Major grading overhaul: difficulty multiplier, tighter scoring, mastery removal, precision penalties

Fix state machine bugs and switch to average scoring for discriminative benchmarking

Fix score aggregation: use max(rewards) for discriminative multi-turn scoring

Remove rate limiter (blocks evaluator) and fix score aggregation to clamped sum

fix(benchmark): Hardening multi-agent environment and strict score compliance

Clamp scores strictly to (0.01, 0.99) to pass OpenEnv Phase 2 continuous environment score verification checks

Fix UI score accumulation logic and save benchmark history

Ensure /step returns info object perfectly matching OpenEnv spec

Fix UI leaderboard fields and commit uv.lock

Fix benchmark output saving: add results dir and print errors

Fix HF_TOKEN parsing for strict validation

first commit