MetaClaw: Just Talk -- An Agent That Meta-Learns and Evolves in the Wild

Abstract

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.

Figures (3)

• Figure 1: Overview of . The framework improves the meta-model $\mathcal{M} = (\theta, \mathcal{S})$ via two complementary loops operating at different timescales. Skill-driven fast adaptation (left) analyzes failed trajectories and instantly expands the skill library $\mathcal{S}$ without parameter updates, taking effect immediately for subsequent tasks. Opportunistic policy optimization (right) accumulates post-adaptation trajectories and, once sufficient data is available, leverages idle signals (sleep, inactivity, calendar) detected by the Opportunistic Meta-Learning Scheduler to trigger RL-based weight updates on $\theta$ via Cloud LoRA fine-tuning.
• Figure 2: Per-day accuracy over 30 simulated workdays (3-day rolling average). Solid lines: GPT-5.2; dashed lines: Kimi-K2.5. MetaClaw (Full) dominates in the mid phase (day 11--22) before difficulty outpaces accumulated knowledge in late days.
• Figure 3: Per-task-type pass rates. File-check (yellow) is unchanged by skills alone but jumps 8.25$\times$ under MetaClaw (Full). Multi-choice (blue) improves with skills but slightly decreases under MetaClaw (Full), reflecting a policy shift toward file-execution.