{"id":"https://openalex.org/W7214720665","doi":"https://doi.org/10.48550/arxiv.2609.33822","title":"Vestrum: Improving Agent Harnesses by Adapting Their Verification, Structure and Memory","display_name":"Vestrum: Improving Agent Harnesses by Adapting Their Verification, Structure and Memory","publication_year":2026,"publication_date":"2026-09-27","ids":{"openalex":"https://openalex.org/W7214720665","doi":"https://doi.org/10.48550/arxiv.2609.33822"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2609.33822","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2609.33822","pdf_url":"https://arxiv.org/pdf/2609.33822","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"volume":null,"type":"preprint","is_traditional":true,"indexed_in":["arxiv","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2609.33822","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5082683981","display_name":"Jayant Parashar","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I165733156","display_name":"University of Georgia","ror":"https://ror.org/00te3t702","country_code":"US","type":"education","lineage":["https://openalex.org/I165733156"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Parashar, Jayant","raw_affiliation_strings":["University of Georgia"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Georgia","institution_ids":["https://openalex.org/I165733156"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5103025096","display_name":"Eugene Douglass","orcid":"https://orcid.org/0000-0001-6225-0329","observed_orcids":["https://orcid.org/0000-0001-6225-0329"]},"institutions":[{"id":"https://openalex.org/I165733156","display_name":"University of Georgia","ror":"https://ror.org/00te3t702","country_code":"US","type":"education","lineage":["https://openalex.org/I165733156"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Douglass, Eugene F.","raw_affiliation_strings":["University of Georgia"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Georgia","institution_ids":["https://openalex.org/I165733156"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5146942361","display_name":"William C Bastian","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I150468666","display_name":"Emory University","ror":"https://ror.org/03czfpz43","country_code":"US","type":"education","lineage":["https://openalex.org/I150468666"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Bastian, William C.","raw_affiliation_strings":["Emory University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Emory University","institution_ids":["https://openalex.org/I150468666"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5111655773","display_name":"Suchendra M. Bhandarkar","orcid":"https://orcid.org/0000-0003-2930-4190","observed_orcids":["https://orcid.org/0000-0003-2930-4190"]},"institutions":[{"id":"https://openalex.org/I165733156","display_name":"University of Georgia","ror":"https://ror.org/00te3t702","country_code":"US","type":"education","lineage":["https://openalex.org/I165733156"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Bhandarkar, Suchendra M.","raw_affiliation_strings":["University of Georgia"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Georgia","institution_ids":["https://openalex.org/I165733156"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":2,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":true,"cited_by_count":0,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":null,"last_page":null},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10906","display_name":"AI-based Problem Solving and Planning","score":0.5845000147819519,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},"topics":[{"id":"https://openalex.org/T10906","display_name":"AI-based Problem Solving and Planning","score":0.5845000147819519,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T10260","display_name":"Software Engineering Research","score":0.06480000168085098,"subfield":{"id":"https://openalex.org/subfields/1710","display_name":"Information Systems"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T13904","display_name":"Artificial Intelligence Applications","score":0.06040000170469284,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/agent-harnesses","display_name":"agent harnesses","score":0.9240000247955322},{"id":"https://openalex.org/keywords/language-model-agents","display_name":"language model agents","score":0.656000018119812},{"id":"https://openalex.org/keywords/failure-analysis","display_name":"failure analysis","score":0.5},{"id":"https://openalex.org/keywords/execution-trace-analysis","display_name":"execution trace analysis","score":0.4909999966621399},{"id":"https://openalex.org/keywords/failure-classification","display_name":"failure classification","score":0.3700000047683716},{"id":"https://openalex.org/keywords/verification","display_name":"verification","score":0.36800000071525574},{"id":"https://openalex.org/keywords/memory-retrieval","display_name":"memory retrieval","score":0.4749999940395355},{"id":"https://openalex.org/keywords/knowledge-synthesis","display_name":"knowledge synthesis","score":0.6359999775886536},{"id":"https://openalex.org/keywords/decomposition","display_name":"decomposition","score":0.34700000286102295},{"id":"https://openalex.org/keywords/evaluation-cost-reduction","display_name":"evaluation cost reduction","score":0.5189999938011169},{"id":"https://openalex.org/keywords/cell-type-identification","display_name":"cell type annotation","score":0.515999972820282}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7975000143051147},{"id":"https://openalex.org/C2780451532","wikidata":"https://www.wikidata.org/wiki/Q759676","display_name":"Task (project management)","level":2,"score":0.642300009727478},{"id":"https://openalex.org/C2778012447","wikidata":"https://www.wikidata.org/wiki/Q1034415","display_name":"Scope (computer science)","level":2,"score":0.5806000232696533},{"id":"https://openalex.org/C2776760102","wikidata":"https://www.wikidata.org/wiki/Q5139990","display_name":"Code (set theory)","level":3,"score":0.49549999833106995},{"id":"https://openalex.org/C2777904410","wikidata":"https://www.wikidata.org/wiki/Q7397","display_name":"Software","level":2,"score":0.48730000853538513},{"id":"https://openalex.org/C199360897","wikidata":"https://www.wikidata.org/wiki/Q9143","display_name":"Programming language","level":1,"score":0.48030000925064087},{"id":"https://openalex.org/C115903868","wikidata":"https://www.wikidata.org/wiki/Q80993","display_name":"Software engineering","level":1,"score":0.45170000195503235},{"id":"https://openalex.org/C2776321320","wikidata":"https://www.wikidata.org/wiki/Q857525","display_name":"Annotation","level":2,"score":0.4316999912261963},{"id":"https://openalex.org/C2777267654","wikidata":"https://www.wikidata.org/wiki/Q3519023","display_name":"Test (biology)","level":2,"score":0.41929998993873596},{"id":"https://openalex.org/C2776175482","wikidata":"https://www.wikidata.org/wiki/Q1195816","display_name":"Transfer (computing)","level":2,"score":0.37860000133514404},{"id":"https://openalex.org/C12186640","wikidata":"https://www.wikidata.org/wiki/Q6815743","display_name":"Memory model","level":3,"score":0.29649999737739563},{"id":"https://openalex.org/C12725497","wikidata":"https://www.wikidata.org/wiki/Q810247","display_name":"Baseline (sea)","level":2,"score":0.296099990606308},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.2809999883174896},{"id":"https://openalex.org/C107457646","wikidata":"https://www.wikidata.org/wiki/Q207434","display_name":"Human\u2013computer interaction","level":1,"score":0.25999999046325684},{"id":"https://openalex.org/C2985957978","wikidata":"https://www.wikidata.org/wiki/Q492","display_name":"Human memory","level":3,"score":0.25949999690055847},{"id":"https://openalex.org/C168167062","wikidata":"https://www.wikidata.org/wiki/Q1117970","display_name":"Component (thermodynamics)","level":2,"score":0.257099986076355},{"id":"https://openalex.org/C45235069","wikidata":"https://www.wikidata.org/wiki/Q278425","display_name":"Table (database)","level":2,"score":0.2533000111579895},{"id":"https://openalex.org/C1009929","wikidata":"https://www.wikidata.org/wiki/Q179550","display_name":"Software bug","level":3,"score":0.2531000077724457},{"id":"https://openalex.org/C12590798","wikidata":"https://www.wikidata.org/wiki/Q3933199","display_name":"Replication (statistics)","level":2,"score":0.2515000104904175},{"id":"https://openalex.org/C25343380","wikidata":"https://www.wikidata.org/wiki/Q277521","display_name":"Relation (database)","level":2,"score":0.2500999867916107}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2609.33822","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2609.33822","pdf_url":"https://arxiv.org/pdf/2609.33822","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2609.33822","doi":"https://doi.org/10.48550/arxiv.2609.33822","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2609.33822","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2609.33822","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2609.33822","pdf_url":"https://arxiv.org/pdf/2609.33822","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7214720665.pdf","grobid_xml":"https://content.openalex.org/works/W7214720665.grobid-xml"},"referenced_works_count":40,"referenced_works":["https://openalex.org/W1555759181","https://openalex.org/W1982723861","https://openalex.org/W4281483047","https://openalex.org/W4310290453","https://openalex.org/W4387430850","https://openalex.org/W4387636003","https://openalex.org/W4399655920","https://openalex.org/W4402684150","https://openalex.org/W4403788984","https://openalex.org/W4404131920","https://openalex.org/W4406840428","https://openalex.org/W4412673546","https://openalex.org/W4415971869","https://openalex.org/W4416437085","https://openalex.org/W4416830533","https://openalex.org/W4416895257","https://openalex.org/W7084066371","https://openalex.org/W7084138381","https://openalex.org/W7106690462","https://openalex.org/W7127393332","https://openalex.org/W7131075195","https://openalex.org/W7131825876","https://openalex.org/W7133222210","https://openalex.org/W7152505967","https://openalex.org/W7153056532","https://openalex.org/W7155622066","https://openalex.org/W7156766069","https://openalex.org/W7158100772","https://openalex.org/W7161237514","https://openalex.org/W7163161372","https://openalex.org/W7163637799","https://openalex.org/W7165761444","https://openalex.org/W7166837933","https://openalex.org/W7166853358","https://openalex.org/W7169879786","https://openalex.org/W7202143511","https://openalex.org/W7202250343","https://openalex.org/W7203715194","https://openalex.org/W7204450245","https://openalex.org/W7213594593"],"related_works":["https://openalex.org/W7204588362","https://openalex.org/W7163637799","https://openalex.org/W7214107704","https://openalex.org/W7215114069","https://openalex.org/W7169541667","https://openalex.org/W7220436573","https://openalex.org/W7160120919","https://openalex.org/W7162690279","https://openalex.org/W7212406378","https://openalex.org/W7220698031"],"abstract_inverted_index":{"An":[0],"agent":[1],"harness":[2,46],"controls":[3],"how":[4],"a":[5,27,36,60,96],"language":[6],"model":[7],"accesses":[8],"information,":[9],"uses":[10],"tools,":[11],"preserves":[12],"memory,":[13],"and":[14,85,88,108,147,165,178],"checks":[15,134],"its":[16],"work.":[17],"Improving":[18],"this":[19],"software":[20],"is":[21,56],"costly":[22],"when":[23],"each":[24],"evaluation":[25],"requires":[26],"long":[28],"interaction":[29],"with":[30],"an":[31],"environment.":[32],"We":[33],"introduce":[34],"Vestrum,":[35],"framework":[37],"that":[38,57,71],"turns":[39],"failures":[40,66,75],"in":[41,172,212],"execution":[42],"traces":[43],"into":[44],"scoped":[45],"changes":[47,80],"without":[48],"training":[49],"the":[50,112,200,208],"task":[51],"model.":[52],"Its":[53],"organizing":[54],"overhypothesis":[55],"tasks":[58,142],"of":[59,133,158],"shared":[61],"kind":[62],"may":[63],"exhibit":[64],"recurring":[65],"whose":[67],"remedies":[68],"transfer":[69],"within":[70],"kind.":[72],"Vestrum":[73,204],"expresses":[74],"as":[76,95],"recognizable":[77],"classes,":[78],"proposes":[79],"across":[81],"verification,":[82],"retrieval,":[83],"decomposition,":[84],"knowledge":[86],"synthesis,":[87],"screens":[89],"their":[90],"scope":[91],"before":[92],"evaluating":[93],"them":[94],"bundle.":[97],"A":[98],"persistent":[99],"lessons":[100],"file":[101],"informs":[102],"subsequent":[103],"proposals.":[104],"Across":[105,167],"five":[106],"settings":[107],"two":[109],"baseline":[110],"harnesses,":[111],"frozen":[113],"harnesses":[114],"improve":[115],"held-out":[116,141,156],"performance:":[117],"UltraHorizon":[118],"rises":[119],"from":[120,129,151],"47.6":[121],"to":[122,131,153,190],"59.8":[123],"over":[124,136],"GAM,":[125],"Terminal-Bench":[126],"4":[127],"Hard":[128],"63.7%":[130],"70.3%":[132],"passed":[135],"Claude":[137],"Code":[138],"on":[139,155,163],"eight":[140],"at":[143,181,184],"1.03x":[144],"test":[145],"cost,":[146],"cell-type":[148],"annotation":[149],"agreement":[150],"67.5%":[152],"77.8%":[154],"sections":[157],"one":[159],"slide,":[160],"alongside":[161],"gains":[162],"LoCoMo":[164],"AMA-Bench.":[166],"our":[168],"searches,":[169],"verification":[170],"grounded":[171],"evidence":[173],"helped":[174],"both":[175],"intermediate":[176,185],"steps":[177],"final":[179],"answers,":[180],"lower":[182],"cost":[183],"steps,":[186],"while":[187],"critics":[188],"asked":[189],"rebuild":[191],"finished":[192],"answers":[193],"broke":[194],"more":[195],"than":[196],"they":[197],"repaired.":[198],"On":[199],"three":[201],"memory":[202],"benchmarks,":[203],"also":[205],"scores":[206],"above":[207],"evaluated":[209],"GEPA":[210],"configurations":[211],"every":[213],"paired":[214],"evaluation.":[215]},"counts_by_year":[],"updated_date":"2026-11-10T11:39:43.021295","created_date":"2026-09-30T00:00:00"}