Skill VM: Write Once, Run Everywhere Efficiently
TL;DR
This work analyzes 118,000 skills and proposes SkillVM, a compilation and runtime system designed for portable and efficient skill execution that achieves up to 3.2× speedup with enhanced parallelism, and 19-50× latency reduction through code solidification.
Abstract
LLM agents increasingly adopt skills as a reusable unit of composition. While skills are shared across diverse agent platforms, current systems treat them as raw context, causing the same skill to behave inconsistently for different agents. This fragility undermines skill portability and execution efficiency. To address this challenge, we analyze 118,000 skills and draw inspiration from traditional compiler design. We treat skills as code and LLMs as heterogeneous processors. To make portability actionable, we decompose a skill's requirements into a set of primitive capabilities, and measure how well each model-harness pair supports them. Based on these capability profiles, we propose SkillVM, a compilation and runtime system designed for portable and efficient skill execution. At compile time, SkillVM performs capability-based compilation, environment binding, and concurrency extraction. At runtime, SkillVM applies JIT code solidification and adaptive recompilation for performance optimization. We evaluate SkillVM across ten LLMs of varying scales and four agent harnesses, covering SkillsBench and representative skill tasks. Results demonstrate that SkillVM significantly improves task completion rates across different models and environments while reducing token consumption by up to 40%. In terms of performance, SkillVM achieves up to 3.2× speedup with enhanced parallelism, and 19-50× latency reduction through code solidification.