MOSCOPT: Mixture-of-Skills Collective Optimization for LLM Agents

arXiv:2609.14399 · cs.AI, cs.CL · Submitted 2026-09-13 · Read on arXiv

cs.AI, cs.CL

Submitted: 2026-09-13

Updated: 2026-09-18

Journal ref: The Pacific Rim International Conference on Artificial Intelligence (PRICAI), 2026

Code: https://github.com/zhangzhenyu13/SummerClaw

License: http://creativecommons.org/licenses/by/4.0/

The gist: Natural language prompts and skills serve as the strategic backbone of LLM-based agents.

Terminology

Abstract

Natural language prompts and skills serve as the strategic backbone of LLM-based agents. Recent advances in prompt and skill optimization have achieved notable gains, yet all existing methods optimize a single text template---missing the synergy among multiple complementary strategies. We propose MOSCOPT, a text-native, parameter-free algorithm that jointly optimizes a pool of N skills and a gating skill G that dynamically selects K skills per step. To effectively optimize the skills, we build the EditAdam with internally maintained dual states. Through the three-phase interleaved updates with EditAdam, the system monotonically improves without gradient or parameter tuning. Extensive experiments and detailed ablations across 5 benchmarks and 3 target LLMs demonstrate that MOSCOPT consistently outperforms all baselines, and confirm that both the mixture-of-skills architecture with selective activation and the collective evolution with three-phase interleaving are essential to its superior performance. Code is released https://github.com/zhangzhenyu13/SummerClaw/tree/master/summerclaw/agent trainer/algorithms/moscopt.

Sources

Related papers