spedrox-sac/soc-marl-env

Fix UI graph rendering logic for CTDE and Skill bonuses

Fix CTDETrainer args

Fix scenario attribute names

Disable use_vllm explicitly in GRPOConfig to stop torch.compile graph node error

Disable torch compile and fast_inference to fix graph node errors

Hardcode Qwen2.5-3B-Instruct model and remove input

Lock model selection to Qwen3-4B

Add openenv-core to Dockerfile dependencies

docs: update Colab notebook and blog links in README

Add project links to README

Fix reward hacking & deploy HF Space Dashboard

feat: implement multi-agent incident response environment with adversarial red teaming and adaptive Theory of Mind coordination

fix: add uv.lock for multi-mode deployment

fix: simplify main() block for openenv validate compatibility

fix: add app_port 8000 for HF Spaces

fix: add HF Spaces YAML frontmatter to README.md

Merge branch 'main' of https://huggingface.co/spaces/spedrox-sac/incident-response-triage

feat: add hackathon-mandated structured logging and validation script

initial commit

docs: remove hackathon-specific metadata and runtime constraints from inference script documentation

Add hackathon inference.py + env var compatibility

Major: Self-evolving RL engine + OpenAI-compatible API simplification

Add BSD 3-Clause License file

Fix NoneType guard in _calculate_investigation_completeness

Major: 6-task environment with advanced grading and reward shaping

Fix agent failures from Haiku + GPT-5-mini real-world runs

Major: ReflAct-based agent architecture with research-backed improvements

Fix NoneType crash: handle None LLM content + add 429 rate limit retry

Fix 422 errors: add stateful HTTP endpoints and action sanitization

Enhance baseline inference script with support for multiple LLM providers and auto-detection of API keys

Add .gitignore, clean up pycache

Initial commit: Incident Response Triage Environment for OpenEnv