RESOURCE2SKILL: Distilling Executable Agent Skills from Human-Created Multimodal Resources
Summary
RESOURCE2SKILL is a framework that distills executable agent skills from multimodal resources like tutorial videos, code repositories, articles, and artifacts into a hierarchical SkillWiki, improving agent performance by 11.9 percentage points over no-skill agents.
View Cached Full Text
Cached at: 07/20/26, 01:25 PM
Paper page - RESOURCE2SKILL: Distilling Executable Agent Skills from Human-Created Multimodal Resources
Source: https://huggingface.co/papers/2606.29538
Abstract
Skillsareausefulabstractionforsoftwareagents,turninghumanandagentexperienceintoreusableproceduralknowledge.Yetexistingskilllibrariesaremostlyhand-written,text-centric,orderivedfromagenttraces,leavingtutorialvideosandothermultimodalhumanresourceslargelyunderused.WepresentRESOURCE2SKILL,aframeworkthatdistillsmultimodalresources,includingtutorialvideos,repositories,articles,andreferenceartifacts,intoexecutableskillsforsoftwareagents.RESOURCE2SKILLorganizestheseskillsasahierarchicalmultimodalSkillWiki,whereeachentrycombinesstructuredtext,code,visualexamples,metadata,andprovenance.Thisdesignpreservescomplementarysignalsfromdifferentresources:videoscapturetemporaloperationsandvisualeffects,codecapturesexecutabletoolpatterns,andarticlesorartifactsprovideconceptualandstylisticgrounding.Atinferencetime,agentsretrieveandcomposerelevantskillsfromthewiki;whencoverageisinsufficient,thesameconstructionoperatorcanacquirenewskillsonline.Acrosssevenpracticalauthoringdomains,RESOURCE2SKILLimprovesaverageoverallscoreby+11.9percentagepointsoverno-skillagentsandoutperformsstrongharnessbaselinesin26of28main-aggregatemodel-domaincells.Ablationsconfirmthevalueofmultimodalskillformat,hierarchicalorganization,sourcediversity,selectionstrategy,andonlineacquisition.
View arXiv pageView PDFProject pageGitHub36Add to collection
Get this paper in your agent:
hf papers read 2606\.29538
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/2606.29538 in a model README.md to link it from this page.
Datasets citing this paper1
#### microsoft/RESOURCE2SKILL Updated3 days ago • 2.05k • 4
Spaces citing this paper0
No Space linking this paper
Cite arxiv.org/abs/2606.29538 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
Search2Skill: Skill Distillation Beyond Knowledge Boundaries Via Rubric-Based Reinforcement Learning
Search2Skill is a framework that trains LLM agents to identify capability gaps, search external sources, and distill retrieved knowledge into reusable skills using rubric-based reinforcement learning, beating baselines on expert-domain benchmarks.
MMG2Skill: Can Agents Distill In-the-Wild Guides into Self-Evolving Skills?
MMG2Skill converts web-based procedural guides into executable skills for agents through closed-loop learning, improving performance across GUI control, gameplay, and card play tasks with macro-average gains of +12.8 to +25.3 percentage points.
WikiSkill: Compiling Agent Experience into Persistent Knowledge for Skill Evolution
WikiSkill introduces a framework that co-evolves agent skills with a persistent knowledge base to systematically accumulate experience and improve performance across models, demonstrating benefits over state-of-the-art skill-evolution methods.
COLLEAGUE.SKILL: Automated AI Skill Generation via Expert Knowledge Distillation
This paper presents COLLEAGUE.SKILL, an open-source system for automatically distilling person-grounded AI skills from heterogeneous traces into inspectable, correctable, and portable skill packages, enabling LLM agents to carry bounded representations of human expertise and interaction style.
SKILL-DISCO: Distilling and Compiling Agent Traces into Reusable Procedural Skills
Skill-DisCo is a framework that distills reusable procedural skills from successful agent traces and compiles them into callable, executable programs. Experiments on ALFWorld and WebArena show improved success rates and reduced agent turns.