Memento-Skills: Let Agents Design Agents

2026-03-19

AI Summary

This work addresses the challenge of enabling general-purpose agents to autonomously design and iteratively optimize task-specific agents without updating the parameters of large foundation models. To this end, the authors propose a memory-based reinforcement learning framework that integrates stateful prompting, a structured skill library, a Read–Write reflection mechanism, and a trainable skill router, establishing a novel end-to-end paradigm for autonomous agent design. This approach facilitates continual learning and capability evolution without requiring model fine-tuning. Empirical evaluations demonstrate significant performance gains, achieving relative accuracy improvements of 26.2% on the General AI Assistants benchmark and 116.2% on Humanity's Last Exam, thereby showcasing the framework’s effectiveness in enhancing agent adaptability and problem-solving proficiency through dynamic skill composition and reflective reasoning.

Technology Category

• Multiagent Systems: Multiagent Learning
• Humans and AI: Human-Aware Planning and Behavior Prediction
• Cognitive Modeling & Cognitive Systems: Agent Architectures

Application Category

• Search and Retrieval-Augmented AI: Agentic search
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Economics, Online Markets and Human Computation: Uses of LLMs and GenAI for marketplace design, bidding, and strategic interactions

Abstract

We introduce \emph{Memento-Skills}, a generalist, continually-learnable LLM agent system that functions as an \emph{agent-designing agent}: it autonomously constructs, adapts, and improves task-specific agents through experience. The system is built on a memory-based reinforcement learning framework with \emph{stateful prompts}, where reusable skills (stored as structured markdown files) serve as persistent, evolving memory. These skills encode both behaviour and context, enabling the agent to carry forward knowledge across interactions. Starting from simple elementary skills (like Web search and terminal operations), the agent continually improves via the \emph{Read--Write Reflective Learning} mechanism introduced in \emph{Memento~2}~\cite{wang2025memento2}. In the \emph{read} phase, a behaviour-trainable skill router selects the most relevant skill conditioned on the current stateful prompt; in the \emph{write} phase, the agent updates and expands its skill library based on new experience. This closed-loop design enables \emph{continual learning without updating LLM parameters}, as all adaptation is realised through the evolution of externalised skills and prompts. Unlike prior approaches that rely on human-designed agents, Memento-Skills enables a generalist agent to \emph{design agents end-to-end} for new tasks. Through iterative skill generation and refinement, the system progressively improves its own capabilities. Experiments on the \emph{General AI Assistants} benchmark and \emph{Humanity's Last Exam} demonstrate sustained gains, achieving 26.2\% and 116.2\% relative improvements in overall accuracy, respectively. Code is available at https://github.com/Memento-Teams/Memento-Skills.

Problem

• agent design
• continual learning
• LLM agent
• skill evolution
• autonomous agent

Innovation

• Memento-Skills
• agent-designing agent
• continual learning
• stateful prompts
• Read–Write Reflective Learning