代码库
ScienceBuddy: Recursive-in-Recursive Self-Improvement for Interactive Scientific Agents
Python
agent-harnessaiforscienceautoresearchllm-agentsrecursive-self-improvement
RLAnything (ICML 2026) & AutoTool (ICML 2026), DemyAgent: Open-Source RL for LLMs and Agentic Scenarios
Python
agent-rlcoding-agententropy-methodgrpogui-agentllm-agentllm-reasoningmulti-agent-reinforcement-learningpporeinforcement-learningrlhf
OpenClaw-RL: Train any agent simply by talking
Python
asynccodinggrpogui-applicationmemory-systemson-policy-distillationopen-clawopenclaw-skillsrlhfsglangskill-learningslimetinker