arXiv:2606.18051v1 Announce Type: new Abstract: LLM agents increasingly rely on external skills -- reusable tool specifications -- but real-world tasks often require composing multiple skills, not just selecting one. We formalize this as the Compositional Skill Routing problem: given a complex user query and a large skill library, decompose the query into atomic sub-tasks, retrieve the appropriate skill for each sub-task, and compose an executable plan. We present SkillWeaver, a decompose-retrieve-compose framework combining an LLM task decomposer, a bi-encoder skill retriever with FAISS index
The rapid advancement of large language models and their increasing use in complex, real-world tasks necessitates sophisticated methods for skill composition and orchestration.
This work addresses a critical bottleneck in the scalability and utility of AI agents, moving beyond single-tool use to enable more complex, multi-step autonomous operations.
The ability of LLM agents to tackle multi-faceted problems will significantly improve, leading to more robust and versatile AI applications in various domains.
- · AI Agent Developers
- · Automation Software Providers
- · Cloud Computing Platforms
- · Enterprises Adopting AI
- · Manual Workflow Providers
- · Legacy Integration Platforms
AI agents become capable of autonomously executing substantially more complex tasks.
Increased efficiency and automation in high-cognitive-load white-collar professions.
Accelerated development of general-purpose AI systems through more sophisticated agentic capabilities.
This signal links to a primary source. Continuum Brief monitors and indexes it as part of the live intelligence stream — we do not republish source content.
Read at arXiv cs.CL