Search papers, labs, and topics across Lattice.
MetaClaw is introduced as a continual meta-learning framework for LLM agents, designed to adapt to evolving user needs without downtime by jointly evolving a base LLM policy and a library of reusable behavioral skills. It uses skill-driven fast adaptation, synthesizing new skills from failure trajectories via an LLM evolver, and opportunistic policy optimization, performing gradient-based updates via cloud LoRA fine-tuning and RL-PRM during inactive periods. Experiments on MetaClaw-Bench and AutoResearchClaw demonstrate that MetaClaw improves accuracy by up to 32% and advances Kimi-K2.5 accuracy from 21.4% to 40.6%.
LLM agents can now learn on the fly and adapt to evolving user needs without disruptive downtime, thanks to a novel meta-learning framework that synthesizes new skills from failure trajectories and optimizes the base policy during inactive periods.
Large language model (LLM) agents are increasingly used for complex tasks, yet deployed agents often remain static, failing to adapt as user needs evolve. This creates a tension between the need for continuous service and the necessity of updating capabilities to match shifting task distributions. On platforms like OpenClaw, which handle diverse workloads across 20+ channels, existing methods either store raw trajectories without distilling knowledge, maintain static skill libraries, or require disruptive downtime for retraining. We present MetaClaw, a continual meta-learning framework that jointly evolves a base LLM policy and a library of reusable behavioral skills. MetaClaw employs two complementary mechanisms. Skill-driven fast adaptation analyzes failure trajectories via an LLM evolver to synthesize new skills, enabling immediate improvement with zero downtime. Opportunistic policy optimization performs gradient-based updates via cloud LoRA fine-tuning and Reinforcement Learning with a Process Reward Model (RL-PRM). This is triggered during user-inactive windows by the Opportunistic Meta-Learning Scheduler (OMLS), which monitors system inactivity and calendar data. These mechanisms are mutually reinforcing: a refined policy generates better trajectories for skill synthesis, while richer skills provide higher-quality data for policy optimization. To prevent data contamination, a versioning mechanism separates support and query data. Built on a proxy-based architecture, MetaClaw scales to production-size LLMs without local GPUs. Experiments on MetaClaw-Bench and AutoResearchClaw show that skill-driven adaptation improves accuracy by up to 32% relative. The full pipeline advances Kimi-K2.5 accuracy from 21.4% to 40.6% and increases composite robustness by 18.3%. Code is available at https://github.com/aiming-lab/MetaClaw.
BAIR