SkillHone: A Harness for Continual Agent Skill Evolution Through Persistent Decision History
Summary
SkillHone is a harness for continual agent skill evolution that uses persistent decision history and practice feedback to improve performance on research and tool-mediated analysis tasks. It outperforms existing methods on GAIA and WebWalkerQA-EN benchmarks.
View Cached Full Text
Cached at: 07/01/26, 11:42 AM
Paper page - SkillHone: A Harness for Continual Agent Skill Evolution Through Persistent Decision History
Source: https://huggingface.co/papers/2606.08671
Abstract
SkillHone enables continuous evolution of agent skills by maintaining persistent decision histories and incorporating practice feedback for improved performance across research and tool-mediated analysis tasks.
Agent skillsextend language-model agents with task-specific procedures, scripts, and references, but the tasks and environments they target continually change. Existing methods improve skills in bounded runs and retain only the final artifact, discarding thedecision historythat later agents need to interpret prior revisions, evaluations, and rejected alternatives. We introduce SkillHone, a harness for continual agentskill evolutiongrounded in persistentdecision history. SkillHone pairs skill revisions with evaluation-side evidence that suppliespractice feedback, recording structured histories of diagnoses, revisions, evidence, and outcomes. Role-separated subagents runcandidate skillson practice probes withredacted reportingand propose revisions informed by prior decisions, enablingcross-session refinementwithout rediscovering past rationale. Ondeep-research benchmarks, SkillHone runs without a pre-integrated search stack and outperforms the commercially backed deep-research agent by 15.8 points onGAIAand 3.2 points onWebWalkerQA-EN, while also exceeding prior skill-evolution methods. We further deploy SkillHone on internaltool-mediated analysisscenarios, where it improves accuracy by an average of 18.8 points across seven settings.
View arXiv pageView PDFProject pageGitHub6Add to collection
Get this paper in your agent:
hf papers read 2606\.08671
Don’t have the latest CLI?curl \-LsSf https://hf\.co/cli/install\.sh \| bash
Models citing this paper0
No model linking this paper
Cite arxiv.org/abs/2606.08671 in a model README.md to link it from this page.
Datasets citing this paper0
No dataset linking this paper
Cite arxiv.org/abs/2606.08671 in a dataset README.md to link it from this page.
Spaces citing this paper0
No Space linking this paper
Cite arxiv.org/abs/2606.08671 in a Space README.md to link it from this page.
Collections including this paper0
No Collection including this paper
Add this paper to acollectionto link it from this page.
Similar Articles
SkillHEX: Improving Agent Skills via Hypothesis-Driven Autonomous Exploration and Exploitation
SkillHEX proposes a closed-loop framework for autonomous skill evolution in LLM agents, using hypothesis-driven self-verification and evidence-guided tree search to overcome sparse reward challenges. It outperforms existing self-evolving methods on SkillsBench with limited interaction budgets.
SkillHarness: Harnessing Safe Skills for Computer-Use Agents
SkillHarness is a framework that enables computer-use agents to safely learn and execute skills in dynamic environments by incorporating safety constraints and adaptive skill selection mechanisms, reducing unsafe rates by 57.1%.
Bayesian-Agent: Posterior-Guided Skill Evolution for LLM Agent Harnesses
Bayesian-Agent presents a framework that treats reusable skills and SOPs as hypotheses, using Bayesian inference to guide agent behavior and improve task performance through posterior-guided harness optimization. It achieves significant improvements on multiple benchmarks with deepseek-v4-flash.
SkillOpt: Executive Strategy for Self-Evolving Agent Skills
SkillOpt introduces a systematic text-space optimizer for agent skills that trains skills as external agent state with stable updates and zero deployment inference overhead, achieving superior performance across multiple benchmarks and execution environments.
SkillFlow:Benchmarking Lifelong Skill Discovery and Evolution for Autonomous Agents
SkillFlow introduces a benchmark of 166 tasks across 20 families for evaluating autonomous agents' ability to discover, repair, and maintain skills over time through a lifelong learning protocol. Experiments reveal a substantial capability gap among leading models, with Claude Opus 4.6 improving significantly while others show limited or negative gains from skill evolution.