MetaCogAgent: A Metacognitive Multi-Agent LLM Framework with Self-Aware Task Delegation

TL;DR

MetaCogAgent introduces a metacognitive multi-agent LLM framework with self-assessment and adaptive task delegation, significantly improving task accuracy and efficiency by enabling agents to evaluate their own capabilities before execution.

Contribution

The paper presents a novel metacognitive framework for multi-agent LLM systems, incorporating self-assessment, adaptive delegation, and capability boundary learning to enhance performance.

Findings

Achieves 82.4% task accuracy, 8.7% above baseline.

Uses 5% fewer API calls than AutoGen.

34% fewer API calls than ensemble voting.

Abstract

Multi-agent large language model (LLM) systems have shown promise for solving complex tasks through agent collaboration. However, existing frameworks assign tasks based on predefined roles without considering whether an agent can accurately assess its own competence boundaries, leading to overconfident execution on tasks beyond its expertise. Inspired by metacognition theory from cognitive science, we propose MetaCogAgent, a multi-agent LLM framework where each agent is equipped with a Metacognitive Self-Assessment Unit that evaluates task-capability alignment before execution. The framework introduces three contributions: (1) a self-assessment mechanism that estimates per-task confidence by combining verbalized uncertainty with historical capability profiles; (2) an adaptive delegation protocol that routes low-confidence tasks to better-suited agents through cross-agent evaluation; and (3) a capability boundary learning module that iteratively refines each agent's competence model via cybernetic feedback. Experiments on our constructed MetaCog-Eval benchmark (700 tasks across 5 cognitive dimensions) demonstrate that MetaCogAgent achieves 82.4% task accuracy -- 8.7% above the best routing baseline -- while using 5% fewer API calls than AutoGen and 34% fewer than ensemble voting. Ablation studies confirm that each metacognitive component contributes to overall system performance.