SkillCombiner: Optimizing Agent Skills via Edit Combination Search

Seungbeom Park; Jaewon Chu; Hyunwoo J. Kim

SkillCombiner: Optimizing Agent Skills via Edit Combination Search

Seungbeom Park, Jaewon Chu, Hyunwoo J. Kim

Correct Metadata for

Use this form to create a GitHub issue with structured data describing the correction. You will need a GitHub account. Once you create that issue, the correction will be reviewed by a staff member.

⚠️ Mobile Users: Submitting this form to create a new issue will only work with github.com, not the GitHub Mobile app.

Important: The Anthology treat PDFs as authoritative. Please use this form only to correct data that is out of line with the PDF. See our corrections guidelines if you need to change the PDF.

Title Adjust the title. Retain tags such as <fixed-case>.

Authors Adjust author names and order to match the PDF.

Abstract Correct abstract if needed. Retain XML formatting tags such as <tex-math>. You may use ... for bold, ... for italic, ... for underline, <sc>...</sc> for small-caps, <tt>...<tt> for typewriter text, <url>...</url> for URLs, <a href=...> for hyperlinks, and <par/> for paragraph breaks.

Verification against PDF Ensure that the new title/authors match the snapshot below. (If there is no snapshot or it is too small, consult the PDF.)

Authors concatenated from the text boxes above:

ALL author names match the snapshot above—including middle initials, hyphens, and accents.

Abstract

Large language model agents are often guided by reusable natural-language skills that provide task-specific instructions. Existing work automatically refines these skills by using LLMs to propose and repeatedly incorporate textual edits from agent feedback. We find that the resulting performance depends substantially on which proposed edits are combined, and adding an edit can even degrade the skill. We therefore introduce SkillCombiner, which generates candidate edits and explicitly searches over their combinations. Across various settings, SkillCombiner matches or exceeds a strong baseline in mean performance while reducing optimizer cost by 75.2% and total optimization-stage cost by 21.5% on average. These results show that edit combination is an important optimization target in agent skill refinement.

Anthology ID:: 2026.realm-1.81
Volume:: Proceedings of the Second Workshop for REsearch on Agent Language Models (REALM 2026)
Month:: October
Year:: 2026
Address:: Budapest, Hungary
Editors:: Ehsan Kamalloo, Nicolas Gontier, Nikolai Rozanov, Maryam Hashemzadeh, Akhil Arora
Venues:: REALM | WS
SIG:
Publisher:: Association for Computational Linguistics
Note:
Pages:: 1188–1197
Language:
URL:: https://aclanthology.org/2026.realm-1.81/
DOI:
Bibkey:
Cite (ACL):: Seungbeom Park, Jaewon Chu, and Hyunwoo J. Kim. 2026. SkillCombiner: Optimizing Agent Skills via Edit Combination Search. In Proceedings of the Second Workshop for REsearch on Agent Language Models (REALM 2026), pages 1188–1197, Budapest, Hungary. Association for Computational Linguistics.
Cite (Informal):: SkillCombiner: Optimizing Agent Skills via Edit Combination Search (Park et al., REALM 2026)
Copy Citation:
PDF:: https://aclanthology.org/2026.realm-1.81.pdf

PDF Cite Search Fix data

Export citation

BibTeX
MODS XML
Endnote
Preformatted

@inproceedings{park-etal-2026-skillcombiner,
    title = "{S}kill{C}ombiner: Optimizing Agent Skills via Edit Combination Search",
    author = "Park, Seungbeom  and
      Chu, Jaewon  and
      Kim, Hyunwoo J.",
    editor = "Kamalloo, Ehsan  and
      Gontier, Nicolas  and
      Rozanov, Nikolai  and
      Hashemzadeh, Maryam  and
      Arora, Akhil",
    booktitle = "Proceedings of the Second Workshop for {RE}search on Agent Language Models ({REALM} 2026)",
    month = oct,
    year = "2026",
    address = "Budapest, Hungary",
    publisher = "Association for Computational Linguistics",
    url = "https://aclanthology.org/2026.realm-1.81/",
    pages = "1188--1197",
    ISBN = "979-8-89176-485-9",
    abstract = "Large language model agents are often guided by reusable natural-language skills that provide task-specific instructions. Existing work automatically refines these skills by using LLMs to propose and repeatedly incorporate textual edits from agent feedback. We find that the resulting performance depends substantially on which proposed edits are combined, and adding an edit can even degrade the skill. We therefore introduce SkillCombiner, which generates candidate edits and explicitly searches over their combinations. Across various settings, SkillCombiner matches or exceeds a strong baseline in mean performance while reducing optimizer cost by 75.2{\%} and total optimization-stage cost by 21.5{\%} on average. These results show that edit combination is an important optimization target in agent skill refinement."
}

Download as File

<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="park-etal-2026-skillcombiner">
    <titleInfo>
        <title>SkillCombiner: Optimizing Agent Skills via Edit Combination Search</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Seungbeom</namePart>
        <namePart type="family">Park</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Jaewon</namePart>
        <namePart type="family">Chu</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Hyunwoo</namePart>
        <namePart type="given">J</namePart>
        <namePart type="family">Kim</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <originInfo>
        <dateIssued>2026-10</dateIssued>
    </originInfo>
    <typeOfResource>text</typeOfResource>
    <relatedItem type="host">
        <titleInfo>
            <title>Proceedings of the Second Workshop for REsearch on Agent Language Models (REALM 2026)</title>
        </titleInfo>
        <name type="personal">
            <namePart type="given">Ehsan</namePart>
            <namePart type="family">Kamalloo</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Nicolas</namePart>
            <namePart type="family">Gontier</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Nikolai</namePart>
            <namePart type="family">Rozanov</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Maryam</namePart>
            <namePart type="family">Hashemzadeh</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Akhil</namePart>
            <namePart type="family">Arora</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <originInfo>
            <publisher>Association for Computational Linguistics</publisher>
            <place>
                <placeTerm type="text">Budapest, Hungary</placeTerm>
            </place>
        </originInfo>
        <genre authority="marcgt">conference publication</genre>
        <identifier type="isbn">979-8-89176-485-9</identifier>
    </relatedItem>
    <abstract>Large language model agents are often guided by reusable natural-language skills that provide task-specific instructions. Existing work automatically refines these skills by using LLMs to propose and repeatedly incorporate textual edits from agent feedback. We find that the resulting performance depends substantially on which proposed edits are combined, and adding an edit can even degrade the skill. We therefore introduce SkillCombiner, which generates candidate edits and explicitly searches over their combinations. Across various settings, SkillCombiner matches or exceeds a strong baseline in mean performance while reducing optimizer cost by 75.2% and total optimization-stage cost by 21.5% on average. These results show that edit combination is an important optimization target in agent skill refinement.</abstract>
    <identifier type="citekey">park-etal-2026-skillcombiner</identifier>
    <location>
        <url>https://aclanthology.org/2026.realm-1.81/</url>
    </location>
    <part>
        <date>2026-10</date>
        <extent unit="page">
            <start>1188</start>
            <end>1197</end>
        </extent>
    </part>
</mods>
</modsCollection>

Download as File

%0 Conference Proceedings
%T SkillCombiner: Optimizing Agent Skills via Edit Combination Search
%A Park, Seungbeom
%A Chu, Jaewon
%A Kim, Hyunwoo J.
%Y Kamalloo, Ehsan
%Y Gontier, Nicolas
%Y Rozanov, Nikolai
%Y Hashemzadeh, Maryam
%Y Arora, Akhil
%S Proceedings of the Second Workshop for REsearch on Agent Language Models (REALM 2026)
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-485-9
%F park-etal-2026-skillcombiner
%X Large language model agents are often guided by reusable natural-language skills that provide task-specific instructions. Existing work automatically refines these skills by using LLMs to propose and repeatedly incorporate textual edits from agent feedback. We find that the resulting performance depends substantially on which proposed edits are combined, and adding an edit can even degrade the skill. We therefore introduce SkillCombiner, which generates candidate edits and explicitly searches over their combinations. Across various settings, SkillCombiner matches or exceeds a strong baseline in mean performance while reducing optimizer cost by 75.2% and total optimization-stage cost by 21.5% on average. These results show that edit combination is an important optimization target in agent skill refinement.
%U https://aclanthology.org/2026.realm-1.81/
%P 1188-1197

Download as File

Markdown (Informal)

[SkillCombiner: Optimizing Agent Skills via Edit Combination Search](https://aclanthology.org/2026.realm-1.81/) (Park et al., REALM 2026)

SkillCombiner: Optimizing Agent Skills via Edit Combination Search (Park et al., REALM 2026)

ACL

Seungbeom Park, Jaewon Chu, and Hyunwoo J. Kim. 2026. SkillCombiner: Optimizing Agent Skills via Edit Combination Search. In Proceedings of the Second Workshop for REsearch on Agent Language Models (REALM 2026), pages 1188–1197, Budapest, Hungary. Association for Computational Linguistics.