Upload folder using huggingface_hub

Updated README, scenarios

Updated models.py

Improved inference.py, scenarios.py and app.py

Improve: multi-step investigation, query_logs, check_threat_intel, run_sandbox actions

updated scenarios

Fix: restore

Fix: clamp cumulative obs.score to (0.01,0.99)

Fix: remove SUMMARY with total_reward > 1 from inference output

Fix: remove SUMMARY with total_reward > 1 from inference output

Force fresh Docker build

Fix: replace all 0.0 rewards with 0.01

Fix: clean 3-task tasks.py no task_critical

Fix: remove task_critical, strictly 3 tasks only

Fix: remove task_critical, keep 3 tasks as required

Fix: clamp env rewards to strictly (0.01, 0.99)

Fix: README

Updated README

Updated README

Fix: task.py

Improvement done: added MITRE ATT&CK scenarios, 4th critical task, ransomware precursor

Fix: import SOCEnvironment directly, no HTTP dependency

Fix: load_dotenv(override=True), os.environ for API vars

Fix: add load_dotenv()

Fix: use HF_TOKEN as per submission checklist

Fix: use os.environ API_KEY first, add score to END line

Fix: API_KEY

Fix: clamp grader scores to strictly (0.01, 0.99)

Fix: health=healthy, add /metadata, /mcp, state in /schema

Fix: main() callable, update openenv.yaml, add ignore files

SOC Incident Response Environment