SkillFlow: Flow-Driven Recursive Skill Evolution for Agentic Orchestration

• URL: http://arxiv.org/abs/2605.14089v1
• Date: Wed, 13 May 2026 20:14:44 GMT
• Title: SkillFlow: Flow-Driven Recursive Skill Evolution for Agentic Orchestration
• Abstract summary: SkillFlow is a flow-based framework that takes a trainable Supervisor as the agent and a structured environment with dynamic skill orchestration.<n>Building on these flow diagnostics, a skill evolution mechanism determines when to evolve, what skills to create or prune, and where decision gaps lie.
• Score: 40.79922760459963
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: In recent years, a variety of powerful LLM-based agentic systems have been applied to automate complex tasks through task orchestration. However, existing orchestration methods still face key challenges, including strategy collapse under reward maximization, high gradient variance with opaque credit assignment, and unguided skill evolution whose decisions are typically made by directly prompting an LLM to judge rather than derived from principled training signals. To address these challenges, we propose SkillFlow, a flow-based framework that takes a trainable Supervisor as the agent and a structured environment with dynamic skill library and frozen executor, automating task orchestration through multi-turn interaction. SkillFlow employs Tempered Trajectory Balance (TTB), a regression-based flow-matching loss that samples trajectories proportional to reward, preserving diverse orchestration strategies rather than collapsing to a single mode. The same flow objective yields a jointly learned backward policy that provides transparent per-step credit assignment at zero additional inference cost. Building on these flow diagnostics, a recursive skill evolution mechanism determines when to evolve, what skills to create or prune, and where decision gaps lie -- closing the loop from training signal to autonomous capability growth. Experimental results on 14 datasets show that SkillFlow significantly outperforms baselines across question answering, mathematical reasoning, code generation, and real-world interactive decision making tasks. Our code is available at https://anonymous.4open.science/r/SkillFlow-E850.

Related papers

• Skill Self-Play: Pushing the Frontier of LLM Capability with Co-Evolving Skills [53.15216425100295]
  We introduce Skill Self-Play (Skill-SP), a co-evolutionary framework comprising a proposer, a solver, and a skill controller.<n>The proposer generates challenging tasks conditioned on dynamically sampled skills; the solver explores candidate solutions to push its capability boundaries; and the skill controller collects execution feedback to update and expand the skill library.<n>This interactive co-evolution effectively bridges the gap between structured verification and open-ended exploration.
  arXiv  Detail & Related papers  (2026-07-24T17:59:22Z)
• Self-evolving LLM agents with in-distribution Optimization [60.05867547965365]
  Large Language Models (LLMs) have recently emerged as powerful controllers for interactive agents in complex environments.<n>We propose Q-Evolve, a self-evolving framework for LLM agents that unifies automatic process-reward labeling and policy learning.<n>We evaluate our method on AlfWorld, WebShop, and ScienceWorld, showing Q-Evolve outperforms strong baselines in sample efficiency, robustness, and overall task performance.
  arXiv  Detail & Related papers  (2026-06-05T15:09:52Z)
• StrLoRA: Towards Streaming Continual Visual Instruction Tuning for MLLMs [9.912407135967836]
  We introduce Streaming CVIT (StrCVIT), where models learn from a stream of data chunks containing a dynamic mixture of tasks.<n>In StrCVIT, a model must simultaneously acquire new abilities, reinforce recurring abilities, and forgetting.<n>We propose StrLoRA, a regularized two-stage expert routing framework.
  arXiv  Detail & Related papers  (2026-05-08T06:16:37Z)
• FlowSteer: Interactive Agentic Workflow Orchestration via End-to-End Reinforcement Learning [49.369614288007334]
  FlowSteer is an end-to-end reinforcement learning framework that takes a lightweight policy model as the agent and an executable canvas environment.<n>We show that FlowSteer significantly outperforms baselines across various tasks.
  arXiv  Detail & Related papers  (2026-02-02T05:30:42Z)
• MermaidFlow: Redefining Agentic Workflow Generation via Safety-Constrained Evolutionary Programming [31.549367981711686]
  We introduce MermaidFlow, a framework that redefines agentic search space through safety-constrained graph evolution.<n>We formulate domain-aware evolutionary operators to preserve semantic correctness while promoting structural diversity.<n>MermaidFlow achieves consistent improvements in success rates and faster convergence to executable plans on the agent reasoning benchmark.
  arXiv  Detail & Related papers  (2025-05-29T01:08:36Z)
• AdaFlow: Imitation Learning with Variance-Adaptive Flow-Based Policies [21.024480978703288]
  We propose AdaFlow, an imitation learning framework based on flow-based generative modeling. AdaFlow represents the policy with state-conditioned ordinary differential equations (ODEs) We show that AdaFlow achieves high performance with fast inference speed.
  arXiv  Detail & Related papers  (2024-02-06T10:15:38Z)
• Distributional GFlowNets with Quantile Flows [73.73721901056662]
  Generative Flow Networks (GFlowNets) are a new family of probabilistic samplers where an agent learns a policy for generating complex structure through a series of decision-making steps. In this work, we adopt a distributional paradigm for GFlowNets, turning each flow function into a distribution, thus providing more informative learning signals during training. Our proposed textitquantile matching GFlowNet learning algorithm is able to learn a risk-sensitive policy, an essential component for handling scenarios with risk uncertainty.
  arXiv  Detail & Related papers  (2023-02-11T22:06:17Z)
• Hierarchical Reinforcement Learning as a Model of Human Task Interleaving [60.95424607008241]
  We develop a hierarchical model of supervisory control driven by reinforcement learning. The model reproduces known empirical effects of task interleaving. The results support hierarchical RL as a plausible model of task interleaving.
  arXiv  Detail & Related papers  (2020-01-04T17:53:28Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.