ClawGymII: Training Models Directly via Black-Box Agent Harnesses
智东西 · wechat · 2026-08-18
Core Concept: Decoupling Execution and Optimization
Renmin University of China proposed ClawGymII, a framework for performing black-box reinforcement learning directly through real Agent Harnesses (e.g., ClaudeCode, OpenClaw).
Key Solutions:
- Scalable Execution: Uses Sandboxes to isolate task rollouts, supporting massive concurrency. Tools are integrated via MCP without modifying Harness internals.
- Trajectory Reconstruction: Addresses fragmented black-box calls with Prefix-Tree Reconstruction, merging shared history and filtering invalid branches (e.g., retries) to form complete training trajectories.
- Mix-Harness Training: Enables a single shared policy model to learn simultaneously from heterogeneous Harnesses like OpenClaw and ClaudeCode.
Results:
- On Qwen3-30B-A3B, black-box RL via OpenClaw and ClaudeCode improved scores on PinchBench by 17.28 and 11.71 points, respectively.
- The framework demonstrated stable improvements across diverse tasks like JobBench and OfficeQA, proving its versatility.
Related event: RUC's ClawGym II Trains Agents via Black-Box RL on Real Harnesses(3 posts)→
More from coding & agent
- Homelab owner lets an AI agent run his network, approving only the 'risky changes' — Rasmic · 2026-10-16
- Maker wires Grok and Claude into ROS2 robot with RealSense stereo cameras — chrismatthieu · 2026-10-16
- CooperBench: AI agents perform 50% worse in teams than working alone — _Hao_Zhu · 2026-10-16
- Meta open-sources Muse Spark Open agent model with day-0 TPU support in SGLang — BanghuaZ · 2026-10-16
- Saying "load-bearing" in your prompt makes Opus 5.5 stop holding back — zacharynado · 2026-10-16
- You.com partners with Paperclip to give every agent live web access — PolarBearby · 2026-10-16