FRAMES: Guarded and Dual-Objective Skill Evolution for Agents in Policy-Governed Enterprise Workflows

2026-08-03

AI Summary

This work addresses the challenges faced by large language model (LLM) agents in policy-constrained enterprise workflows—such as document auditing—including sparse feedback, performance degradation due to frequent rule updates, and the need to balance accuracy, reasoning cost, and auditability. To tackle these issues, the authors propose FRAMES, a framework that bootstraps deployable skills during cold-start and enables continuous skill evolution through consensus-based skill mutation, Pareto-optimal trade-offs between precision and computational cost, and a degradation-aware validation mechanism within a closed-loop system. FRAMES is the first approach to achieve auditable, efficient, and robust skill iteration in policy-intensive settings, demonstrating state-of-the-art precision–cost trade-offs on both an internal production system and the tau-bench benchmark, significantly outperforming existing baselines.

Abstract

LLM agents increasingly run policy-bound enterprise workflows such as document auditing, where they must apply rules consistently, ground every value, and stay auditable. Improving these agents is hard: operational feedback is sparse and unlabeled, edits to one rule can regress unrelated cases, and accuracy must improve without inflating inference cost or losing auditability. We present FRAMES, a closed-loop framework that cold-starts deployable skills from existing assets and then evolves them through consensus-based mutation, Pareto selection over accuracy and cost, and an anti-regression guarantee, all while preserving auditability. Deployed on our internal production system, FRAMES attains the best accuracy-cost trade-off among baselines, with the same gains reproduced on tau-bench.

Problem

• policy-governed workflows
• LLM agents
• auditability
• accuracy-cost trade-off
• regression prevention

Innovation

• skill evolution
• policy-governed workflows
• Pareto optimization
• anti-regression guarantee
• auditability