{"id":"4e1ab952f26a86686b8ddce532ab9865f4d41f63470c8050793a36a0aaea1e37","title":"arXiv paper: AgentSelect benchmark for query-to-agent recommendation","argument":"The authors of AgentSelect argue that choosing an AI agent for a particular request is difficult because existing leaderboards and benchmarks assess components separately and use fragmented tasks, metrics, and candidate pools. They present a benchmark for query-to-agent recommendation: matching a narrative request with an agent configuration that combines a backbone model and a toolkit. The benchmark brings together 111,179 queries, 107,721 deployable agents, and 251,103 interaction records from more than 40 sources. The authors report that recommendation data is concentrated in a long tail of nearly one-off examples, making popularity-based CF/GNN methods fragile and capability-aware matching important. They also report that synthesized compositional interactions can be learned, and that models trained on AgentSelect gain on an unseen public agent marketplace, MuleRun. These results make the benchmark a proposed way to study recommendation across a varied agent ecosystem, but the abstract does not establish how well its reported gains carry over to other settings.","topicIds":["intelligence-and-discovery"],"domain":"science","claimType":"attributed reading","limits":"The evidence provided is the arXiv abstract, not the paper's full methods or results; it does not give details that would independently assess the reported performance or transfer gains. The abstract reports results from the authors; those claims are not independently verified by this source page.","speaker":{"name":"Shi, Yunxiao, Xu, Wujiang, Chen, Tingwei, Shang, Haoning, Yang, Ling, Wan, Yunfeng, Cao, Zhuo, Zi, Xing, Metaxas, Dimitris N., Xu, Min","roleAtSource":"Article author"},"source":{"title":"AgentSelect: Benchmark for Narrative Query-to-Agent Recommendation","publisher":"arXiv cs.AI","format":"Article","url":"https://arxiv.org/abs/2603.03761","publishedAt":"2026-03-04","dateBasis":"Date basis not provided."},"evidence":[{"text":"The authors frame AgentSelect as addressing a gap in query-conditioned supervision for recommending end-to-end agent configurations.","excerpt":"there is little query-conditioned supervision for learning to recommend end-to-end agent configurations","url":"https://arxiv.org/abs/2603.03761","relationship":"Source-stated"},{"text":"AgentSelect combines queries, deployable agents, and interaction records aggregated from over 40 sources.","excerpt":"111,179 queries, 107,721 deployable agents, and 251,103 interaction records aggregated from 40+ sources","url":"https://arxiv.org/abs/2603.03761","relationship":"Source-stated"},{"text":"The authors report that long-tail, near one-off supervision makes popularity-based CF/GNN methods fragile and capability-aware matching essential.","excerpt":"a regime shift from dense head reuse to long-tail, near one-off supervision, where popularity-based CF/GNN methods become fragile and content-aware capability matching is essential","url":"https://arxiv.org/abs/2603.03761","relationship":"Source-stated"},{"text":"The authors report that models trained on AgentSelect transfer to the public agent marketplace MuleRun, with consistent gains on an unseen catalog.","excerpt":"models trained on AgentSelect also transfer to a public agent marketplace (MuleRun), yielding consistent gains on an unseen catalog","url":"https://arxiv.org/abs/2603.03761","relationship":"Source-stated"}],"sourceMedia":{"status":"not_found","sourceUrl":"https://arxiv.org/abs/2603.03761","checkedAt":"2026-11-08T21:48:07.989Z","items":[],"note":"No suitable media was found by the supported page extraction."},"assumptions":[],"openQuestion":{"attribution":"Open question","text":"How do the reported gains on MuleRun vary across the unseen catalog, and what evaluation evidence supports their consistency?"},"whatWouldChangeView":{"attribution":"Change condition not recorded","text":""},"editorialComparisons":[],"live":{"title":"arXiv paper: AgentSelect benchmark for query-to-agent recommendation","summary":"The authors of AgentSelect argue that choosing an AI agent for a particular request is difficult because existing leaderboards and benchmarks assess components separately and use fragmented tasks, metrics, and candidate pools. They present a benchmark for query-to-agent recommendation: matching a narrative request with an agent configuration that combines a backbone model and a toolkit. The benchmark brings together 111,179 queries, 107,721 deployable agents, and 251,103 interaction records from more than 40 sources. The authors report that recommendation data is concentrated in a long tail of nearly one-off examples, making popularity-based CF/GNN methods fragile and capability-aware matching important. They also report that synthesized compositional interactions can be learned, and that models trained on AgentSelect gain on an unseen public agent marketplace, MuleRun. These results make the benchmark a proposed way to study recommendation across a varied agent ecosystem, but the abstract does not establish how well its reported gains carry over to other settings.","question":"How do the reported gains on MuleRun vary across the unseen catalog, and what evaluation evidence supports their consistency?","claims":[{"text":"The authors frame AgentSelect as addressing a gap in query-conditioned supervision for recommending end-to-end agent configurations.","excerpt":"there is little query-conditioned supervision for learning to recommend end-to-end agent configurations","support":"source-stated"},{"text":"AgentSelect combines queries, deployable agents, and interaction records aggregated from over 40 sources.","excerpt":"111,179 queries, 107,721 deployable agents, and 251,103 interaction records aggregated from 40+ sources","support":"source-stated"},{"text":"The authors report that long-tail, near one-off supervision makes popularity-based CF/GNN methods fragile and capability-aware matching essential.","excerpt":"a regime shift from dense head reuse to long-tail, near one-off supervision, where popularity-based CF/GNN methods become fragile and content-aware capability matching is essential","support":"source-stated"},{"text":"The authors report that models trained on AgentSelect transfer to the public agent marketplace MuleRun, with consistent gains on an unseen catalog.","excerpt":"models trained on AgentSelect also transfer to a public agent marketplace (MuleRun), yielding consistent gains on an unseen catalog","support":"source-stated"}],"limitations":["The evidence provided is the arXiv abstract, not the paper's full methods or results; it does not give details that would independently assess the reported performance or transfer gains.","The abstract reports results from the authors; those claims are not independently verified by this source page."],"mediaId":"0b898d5968d050bfa71db959e416f2b6b0b7a878a52d6af960bc896cfaacd6b0","correctionReason":null,"id":"4e1ab952f26a86686b8ddce532ab9865f4d41f63470c8050793a36a0aaea1e37","publicationId":"9ac46481-dce2-4edc-8c03-27fe6d25d5cf","version":1,"sourceUrl":"https://arxiv.org/abs/2603.03761","sourceTitle":"AgentSelect: Benchmark for Narrative Query-to-Agent Recommendation","publisher":"arXiv cs.AI","publishedAt":null,"originalPublished":{"value":"2026-03-04","precision":"day","basis":"publisher_citation_date"},"firstObservedAt":"2026-11-06T05:00:50.424Z","createdAt":"2026-11-08T21:17:05.469Z","historical":true,"entityIds":["arxiv"],"topicIds":["intelligence-and-discovery"],"media":[{"id":"0b898d5968d050bfa71db959e416f2b6b0b7a878a52d6af960bc896cfaacd6b0","kind":"article","url":"https://arxiv.org/abs/2603.03761","sourceUrl":"https://arxiv.org/abs/2603.03761","label":"Read original","provenance":"source_link","verification":{"status":"verified","checkedAt":"2026-11-08T21:16:17.691Z","method":"acquired_source","contentType":"text/html"},"recommendationReason":"Recommended by Jev for understanding this source; checked by Luna."}],"sourceMedia":{"status":"not_found","sourceUrl":"https://arxiv.org/abs/2603.03761","checkedAt":"2026-11-08T21:48:07.989Z","items":[],"note":"No suitable media was found by the supported page extraction."},"editorialDecision":{"outcome":"accept","reasonCodes":["material_development"],"rubricVersion":"stream-admission-v5","mediaId":"0b898d5968d050bfa71db959e416f2b6b0b7a878a52d6af960bc896cfaacd6b0","confidence":0.8400000000000001,"provider":"typesafe","model":"jev-1.13.0","inputHash":"a1bcc183720e222dcd6ec8fb2d2571b89caf60fcfc31a40cee90be1916356a66","decidedAt":"2026-11-08T21:16:18.559Z"},"provenance":{"textHash":"ab0c3465e9a1e0bdaadbd5ecdc8875df8a0939f122deb944eeac559d50b66b53","observationId":"170000c695d08d634409fa270f1ec8aab2434d9846801875e124cbd4d8ea5471","retrievedAt":"2026-11-08T21:16:17.568Z","kind":"publisher-page","preparer":{"model":"gpt-6-luna","runId":"36a8a050-fdec-471f-8212-50c051230930"},"reviewer":{"model":"gpt-6-luna","runId":"f473c3ce-6ba5-4653-8642-fd3ce45c36a5"}},"audit":{"status":"uncertain","reviewedAt":"2026-11-08T21:17:05.469Z","model":"anthropic/claude-opus-5.5"},"sourceAttribution":{"authors":[{"name":"Shi, Yunxiao","url":null,"kind":"unknown"},{"name":"Xu, Wujiang","url":null,"kind":"unknown"},{"name":"Chen, Tingwei","url":null,"kind":"unknown"},{"name":"Shang, Haoning","url":null,"kind":"unknown"},{"name":"Yang, Ling","url":null,"kind":"unknown"},{"name":"Wan, Yunfeng","url":null,"kind":"unknown"},{"name":"Cao, Zhuo","url":null,"kind":"unknown"},{"name":"Zi, Xing","url":null,"kind":"unknown"},{"name":"Metaxas, Dimitris N.","url":null,"kind":"unknown"},{"name":"Xu, Min","url":null,"kind":"unknown"}],"authorRole":"article-author","hosts":[],"status":"stated","authorRoster":{"truncated":false,"omitted":0,"assertionGroupsOmitted":0},"participation":[{"actor":{"name":"Shi, Yunxiao","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Shi, Yunxiao","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2603.03761","locator":"head/meta[name=author|citation_author]"},"assertionId":"f11c3ada3041d31719c30c203e2d809ecaf7692749e6511c4699584e5c6c8d3d:participation:0","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-08T21:48:07.989Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"f11c3ada3041d31719c30c203e2d809ecaf7692749e6511c4699584e5c6c8d3d","sourceUrl":"https://arxiv.org/abs/2603.03761","contentSha256":"07395af2ff8798b35c8b7f5ab3f4d462058822cd14ee4e7d8e6340f236e6b673","textHash":"ab0c3465e9a1e0bdaadbd5ecdc8875df8a0939f122deb944eeac559d50b66b53","observedAt":"2026-11-08T21:48:07.989Z"}},{"actor":{"name":"Xu, Wujiang","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Xu, Wujiang","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2603.03761","locator":"head/meta[name=author|citation_author]"},"assertionId":"f11c3ada3041d31719c30c203e2d809ecaf7692749e6511c4699584e5c6c8d3d:participation:1","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-08T21:48:07.989Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"f11c3ada3041d31719c30c203e2d809ecaf7692749e6511c4699584e5c6c8d3d","sourceUrl":"https://arxiv.org/abs/2603.03761","contentSha256":"07395af2ff8798b35c8b7f5ab3f4d462058822cd14ee4e7d8e6340f236e6b673","textHash":"ab0c3465e9a1e0bdaadbd5ecdc8875df8a0939f122deb944eeac559d50b66b53","observedAt":"2026-11-08T21:48:07.989Z"}},{"actor":{"name":"Chen, Tingwei","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Chen, Tingwei","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2603.03761","locator":"head/meta[name=author|citation_author]"},"assertionId":"f11c3ada3041d31719c30c203e2d809ecaf7692749e6511c4699584e5c6c8d3d:participation:2","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-08T21:48:07.989Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"f11c3ada3041d31719c30c203e2d809ecaf7692749e6511c4699584e5c6c8d3d","sourceUrl":"https://arxiv.org/abs/2603.03761","contentSha256":"07395af2ff8798b35c8b7f5ab3f4d462058822cd14ee4e7d8e6340f236e6b673","textHash":"ab0c3465e9a1e0bdaadbd5ecdc8875df8a0939f122deb944eeac559d50b66b53","observedAt":"2026-11-08T21:48:07.989Z"}},{"actor":{"name":"Shang, Haoning","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Shang, Haoning","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2603.03761","locator":"head/meta[name=author|citation_author]"},"assertionId":"f11c3ada3041d31719c30c203e2d809ecaf7692749e6511c4699584e5c6c8d3d:participation:3","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-08T21:48:07.989Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"f11c3ada3041d31719c30c203e2d809ecaf7692749e6511c4699584e5c6c8d3d","sourceUrl":"https://arxiv.org/abs/2603.03761","contentSha256":"07395af2ff8798b35c8b7f5ab3f4d462058822cd14ee4e7d8e6340f236e6b673","textHash":"ab0c3465e9a1e0bdaadbd5ecdc8875df8a0939f122deb944eeac559d50b66b53","observedAt":"2026-11-08T21:48:07.989Z"}},{"actor":{"name":"Yang, Ling","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Yang, Ling","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2603.03761","locator":"head/meta[name=author|citation_author]"},"assertionId":"f11c3ada3041d31719c30c203e2d809ecaf7692749e6511c4699584e5c6c8d3d:participation:4","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-08T21:48:07.989Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"f11c3ada3041d31719c30c203e2d809ecaf7692749e6511c4699584e5c6c8d3d","sourceUrl":"https://arxiv.org/abs/2603.03761","contentSha256":"07395af2ff8798b35c8b7f5ab3f4d462058822cd14ee4e7d8e6340f236e6b673","textHash":"ab0c3465e9a1e0bdaadbd5ecdc8875df8a0939f122deb944eeac559d50b66b53","observedAt":"2026-11-08T21:48:07.989Z"}},{"actor":{"name":"Wan, Yunfeng","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Wan, Yunfeng","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2603.03761","locator":"head/meta[name=author|citation_author]"},"assertionId":"f11c3ada3041d31719c30c203e2d809ecaf7692749e6511c4699584e5c6c8d3d:participation:5","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-08T21:48:07.989Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"f11c3ada3041d31719c30c203e2d809ecaf7692749e6511c4699584e5c6c8d3d","sourceUrl":"https://arxiv.org/abs/2603.03761","contentSha256":"07395af2ff8798b35c8b7f5ab3f4d462058822cd14ee4e7d8e6340f236e6b673","textHash":"ab0c3465e9a1e0bdaadbd5ecdc8875df8a0939f122deb944eeac559d50b66b53","observedAt":"2026-11-08T21:48:07.989Z"}},{"actor":{"name":"Cao, Zhuo","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Cao, Zhuo","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2603.03761","locator":"head/meta[name=author|citation_author]"},"assertionId":"f11c3ada3041d31719c30c203e2d809ecaf7692749e6511c4699584e5c6c8d3d:participation:6","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-08T21:48:07.989Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"f11c3ada3041d31719c30c203e2d809ecaf7692749e6511c4699584e5c6c8d3d","sourceUrl":"https://arxiv.org/abs/2603.03761","contentSha256":"07395af2ff8798b35c8b7f5ab3f4d462058822cd14ee4e7d8e6340f236e6b673","textHash":"ab0c3465e9a1e0bdaadbd5ecdc8875df8a0939f122deb944eeac559d50b66b53","observedAt":"2026-11-08T21:48:07.989Z"}},{"actor":{"name":"Zi, Xing","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Zi, Xing","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2603.03761","locator":"head/meta[name=author|citation_author]"},"assertionId":"f11c3ada3041d31719c30c203e2d809ecaf7692749e6511c4699584e5c6c8d3d:participation:7","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-08T21:48:07.989Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"f11c3ada3041d31719c30c203e2d809ecaf7692749e6511c4699584e5c6c8d3d","sourceUrl":"https://arxiv.org/abs/2603.03761","contentSha256":"07395af2ff8798b35c8b7f5ab3f4d462058822cd14ee4e7d8e6340f236e6b673","textHash":"ab0c3465e9a1e0bdaadbd5ecdc8875df8a0939f122deb944eeac559d50b66b53","observedAt":"2026-11-08T21:48:07.989Z"}},{"actor":{"name":"Metaxas, Dimitris N.","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Metaxas, Dimitris N.","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2603.03761","locator":"head/meta[name=author|citation_author]"},"assertionId":"f11c3ada3041d31719c30c203e2d809ecaf7692749e6511c4699584e5c6c8d3d:participation:8","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-08T21:48:07.989Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"f11c3ada3041d31719c30c203e2d809ecaf7692749e6511c4699584e5c6c8d3d","sourceUrl":"https://arxiv.org/abs/2603.03761","contentSha256":"07395af2ff8798b35c8b7f5ab3f4d462058822cd14ee4e7d8e6340f236e6b673","textHash":"ab0c3465e9a1e0bdaadbd5ecdc8875df8a0939f122deb944eeac559d50b66b53","observedAt":"2026-11-08T21:48:07.989Z"}},{"actor":{"name":"Xu, Min","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Xu, Min","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2603.03761","locator":"head/meta[name=author|citation_author]"},"assertionId":"f11c3ada3041d31719c30c203e2d809ecaf7692749e6511c4699584e5c6c8d3d:participation:9","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-08T21:48:07.989Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"f11c3ada3041d31719c30c203e2d809ecaf7692749e6511c4699584e5c6c8d3d","sourceUrl":"https://arxiv.org/abs/2603.03761","contentSha256":"07395af2ff8798b35c8b7f5ab3f4d462058822cd14ee4e7d8e6340f236e6b673","textHash":"ab0c3465e9a1e0bdaadbd5ecdc8875df8a0939f122deb944eeac559d50b66b53","observedAt":"2026-11-08T21:48:07.989Z"}}],"participationCoverage":{"total":10,"returned":10,"complete":true,"fullAssertions":"source-metadata-endpoint"},"distribution":null,"publisher":null,"sourceUrl":"https://arxiv.org/abs/2603.03761","snapshotId":"f11c3ada3041d31719c30c203e2d809ecaf7692749e6511c4699584e5c6c8d3d","contentSha256":"07395af2ff8798b35c8b7f5ab3f4d462058822cd14ee4e7d8e6340f236e6b673","textHash":"ab0c3465e9a1e0bdaadbd5ecdc8875df8a0939f122deb944eeac559d50b66b53","observedAt":"2026-11-08T21:48:07.989Z","basis":"publisher-page-metadata","sourceLinks":51,"linksTruncated":false,"metadataUrl":"/api/pipeline/stream/4e1ab952f26a86686b8ddce532ab9865f4d41f63470c8050793a36a0aaea1e37/source"}},"canonicalUrl":"https://escapemeridian.com/readings/arxiv-paper-agentselect-benchmark-for-query-to-agent-4e1ab952","formats":{"html":"/readings/arxiv-paper-agentselect-benchmark-for-query-to-agent-4e1ab952","markdown":"/readings/arxiv-paper-agentselect-benchmark-for-query-to-agent-4e1ab952.md","json":"/readings/arxiv-paper-agentselect-benchmark-for-query-to-agent-4e1ab952.json"},"attribution":"Headline and summary by Escape Meridian, prepared by Luna from the attributed source. A source check is not independent confirmation of the source’s claims.","maintainedGuideLinks":[],"readerContext":{"version":"reader-context-v1","attribution":"Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement.","terms":[{"id":"term-benchmark","name":"Benchmark","explanation":"A defined set of tests used to compare systems. Results depend on the tasks and measures selected.","referenceUrl":"https://arxiv.org/abs/2211.09110"},{"id":"term-ai-agent","name":"AI agent","explanation":"An AI system that chooses and carries out actions toward a goal, often using software tools.","referenceUrl":"https://www.ibm.com/think/topics/ai-agents"}],"entities":[{"id":"arxiv","name":"arXiv","kind":"organization","context":"arXiv is a research manuscript repository where authors share papers that may not yet have been peer reviewed.","referenceUrls":["https://arxiv.org/","https://info.arxiv.org/help/api/index.html"],"path":"/entities/arxiv","json":"/entities/arxiv.json","markdown":"/entities/arxiv.md"},{"id":"marketplace","name":"Marketplace","kind":"organization","context":"Marketplace is a public radio business news outlet publishing reporting on the economy, companies, technology and work.","referenceUrls":["https://www.marketplace.org/"],"path":"/entities/marketplace","json":"/entities/marketplace.json","markdown":"/entities/marketplace.md"}]},"researchQuestions":[],"withdrawn":false,"references":{"schemaVersion":1,"target":{"type":"reading","id":"4e1ab952f26a86686b8ddce532ab9865f4d41f63470c8050793a36a0aaea1e37","revision":"9ac46481-dce2-4edc-8c03-27fe6d25d5cf","revisionKind":"publication-id","title":"arXiv paper: AgentSelect benchmark for query-to-agent recommendation","url":"https://escapemeridian.com/readings/arxiv-paper-agentselect-benchmark-for-query-to-agent-4e1ab952"},"source":{"type":"source","id":"https://arxiv.org/abs/2603.03761","revision":"ab0c3465e9a1e0bdaadbd5ecdc8875df8a0939f122deb944eeac559d50b66b53","revisionKind":"sha256-source-text","title":"AgentSelect: Benchmark for Narrative Query-to-Agent Recommendation","url":"https://arxiv.org/abs/2603.03761"},"status":"available","items":[],"next":null,"scope":"Explicit citations in public readings and public file attachments to this reading. Shared topics are not citations. Other artifact types need an explicit reference adapter."},"revisions":[{"id":"4e1ab952f26a86686b8ddce532ab9865f4d41f63470c8050793a36a0aaea1e37","title":"arXiv paper: AgentSelect benchmark for query-to-agent recommendation","argument":"The authors of AgentSelect argue that choosing an AI agent for a particular request is difficult because existing leaderboards and benchmarks assess components separately and use fragmented tasks, metrics, and candidate pools. They present a benchmark for query-to-agent recommendation: matching a narrative request with an agent configuration that combines a backbone model and a toolkit. The benchmark brings together 111,179 queries, 107,721 deployable agents, and 251,103 interaction records from more than 40 sources. The authors report that recommendation data is concentrated in a long tail of nearly one-off examples, making popularity-based CF/GNN methods fragile and capability-aware matching important. They also report that synthesized compositional interactions can be learned, and that models trained on AgentSelect gain on an unseen public agent marketplace, MuleRun. These results make the benchmark a proposed way to study recommendation across a varied agent ecosystem, but the abstract does not establish how well its reported gains carry over to other settings.","topicIds":["intelligence-and-discovery"],"domain":"science","claimType":"attributed reading","limits":"The evidence provided is the arXiv abstract, not the paper's full methods or results; it does not give details that would independently assess the reported performance or transfer gains. The abstract reports results from the authors; those claims are not independently verified by this source page.","speaker":{"name":"Shi, Yunxiao, Xu, Wujiang, Chen, Tingwei, Shang, Haoning, Yang, Ling, Wan, Yunfeng, Cao, Zhuo, Zi, Xing, Metaxas, Dimitris N., Xu, Min","roleAtSource":"Article author"},"source":{"title":"AgentSelect: Benchmark for Narrative Query-to-Agent Recommendation","publisher":"arXiv cs.AI","format":"Article","url":"https://arxiv.org/abs/2603.03761","publishedAt":"2026-03-04","dateBasis":"Date basis not provided."},"evidence":[{"text":"The authors frame AgentSelect as addressing a gap in query-conditioned supervision for recommending end-to-end agent configurations.","excerpt":"there is little query-conditioned supervision for learning to recommend end-to-end agent configurations","url":"https://arxiv.org/abs/2603.03761","relationship":"Source-stated"},{"text":"AgentSelect combines queries, deployable agents, and interaction records aggregated from over 40 sources.","excerpt":"111,179 queries, 107,721 deployable agents, and 251,103 interaction records aggregated from 40+ sources","url":"https://arxiv.org/abs/2603.03761","relationship":"Source-stated"},{"text":"The authors report that long-tail, near one-off supervision makes popularity-based CF/GNN methods fragile and capability-aware matching essential.","excerpt":"a regime shift from dense head reuse to long-tail, near one-off supervision, where popularity-based CF/GNN methods become fragile and content-aware capability matching is essential","url":"https://arxiv.org/abs/2603.03761","relationship":"Source-stated"},{"text":"The authors report that models trained on AgentSelect transfer to the public agent marketplace MuleRun, with consistent gains on an unseen catalog.","excerpt":"models trained on AgentSelect also transfer to a public agent marketplace (MuleRun), yielding consistent gains on an unseen catalog","url":"https://arxiv.org/abs/2603.03761","relationship":"Source-stated"}],"sourceMedia":{"status":"not_found","sourceUrl":"https://arxiv.org/abs/2603.03761","checkedAt":"2026-11-08T21:48:07.989Z","items":[],"note":"No suitable media was found by the supported page extraction."},"assumptions":[],"openQuestion":{"attribution":"Open question","text":"How do the reported gains on MuleRun vary across the unseen catalog, and what evaluation evidence supports their consistency?"},"whatWouldChangeView":{"attribution":"Change condition not recorded","text":""},"editorialComparisons":[],"live":{"title":"arXiv paper: AgentSelect benchmark for query-to-agent recommendation","summary":"The authors of AgentSelect argue that choosing an AI agent for a particular request is difficult because existing leaderboards and benchmarks assess components separately and use fragmented tasks, metrics, and candidate pools. They present a benchmark for query-to-agent recommendation: matching a narrative request with an agent configuration that combines a backbone model and a toolkit. The benchmark brings together 111,179 queries, 107,721 deployable agents, and 251,103 interaction records from more than 40 sources. The authors report that recommendation data is concentrated in a long tail of nearly one-off examples, making popularity-based CF/GNN methods fragile and capability-aware matching important. They also report that synthesized compositional interactions can be learned, and that models trained on AgentSelect gain on an unseen public agent marketplace, MuleRun. These results make the benchmark a proposed way to study recommendation across a varied agent ecosystem, but the abstract does not establish how well its reported gains carry over to other settings.","question":"How do the reported gains on MuleRun vary across the unseen catalog, and what evaluation evidence supports their consistency?","claims":[{"text":"The authors frame AgentSelect as addressing a gap in query-conditioned supervision for recommending end-to-end agent configurations.","excerpt":"there is little query-conditioned supervision for learning to recommend end-to-end agent configurations","support":"source-stated"},{"text":"AgentSelect combines queries, deployable agents, and interaction records aggregated from over 40 sources.","excerpt":"111,179 queries, 107,721 deployable agents, and 251,103 interaction records aggregated from 40+ sources","support":"source-stated"},{"text":"The authors report that long-tail, near one-off supervision makes popularity-based CF/GNN methods fragile and capability-aware matching essential.","excerpt":"a regime shift from dense head reuse to long-tail, near one-off supervision, where popularity-based CF/GNN methods become fragile and content-aware capability matching is essential","support":"source-stated"},{"text":"The authors report that models trained on AgentSelect transfer to the public agent marketplace MuleRun, with consistent gains on an unseen catalog.","excerpt":"models trained on AgentSelect also transfer to a public agent marketplace (MuleRun), yielding consistent gains on an unseen catalog","support":"source-stated"}],"limitations":["The evidence provided is the arXiv abstract, not the paper's full methods or results; it does not give details that would independently assess the reported performance or transfer gains.","The abstract reports results from the authors; those claims are not independently verified by this source page."],"mediaId":"0b898d5968d050bfa71db959e416f2b6b0b7a878a52d6af960bc896cfaacd6b0","correctionReason":null,"id":"4e1ab952f26a86686b8ddce532ab9865f4d41f63470c8050793a36a0aaea1e37","publicationId":"9ac46481-dce2-4edc-8c03-27fe6d25d5cf","version":1,"sourceUrl":"https://arxiv.org/abs/2603.03761","sourceTitle":"AgentSelect: Benchmark for Narrative Query-to-Agent Recommendation","publisher":"arXiv cs.AI","publishedAt":null,"originalPublished":{"value":"2026-03-04","precision":"day","basis":"publisher_citation_date"},"firstObservedAt":"2026-11-06T05:00:50.424Z","createdAt":"2026-11-08T21:17:05.469Z","historical":true,"entityIds":["arxiv"],"topicIds":["intelligence-and-discovery"],"media":[{"id":"0b898d5968d050bfa71db959e416f2b6b0b7a878a52d6af960bc896cfaacd6b0","kind":"article","url":"https://arxiv.org/abs/2603.03761","sourceUrl":"https://arxiv.org/abs/2603.03761","label":"Read original","provenance":"source_link","verification":{"status":"verified","checkedAt":"2026-11-08T21:16:17.691Z","method":"acquired_source","contentType":"text/html"},"recommendationReason":"Recommended by Jev for understanding this source; checked by Luna."}],"sourceMedia":{"status":"not_found","sourceUrl":"https://arxiv.org/abs/2603.03761","checkedAt":"2026-11-08T21:48:07.989Z","items":[],"note":"No suitable media was found by the supported page extraction."},"editorialDecision":{"outcome":"accept","reasonCodes":["material_development"],"rubricVersion":"stream-admission-v5","mediaId":"0b898d5968d050bfa71db959e416f2b6b0b7a878a52d6af960bc896cfaacd6b0","confidence":0.8400000000000001,"provider":"typesafe","model":"jev-1.13.0","inputHash":"a1bcc183720e222dcd6ec8fb2d2571b89caf60fcfc31a40cee90be1916356a66","decidedAt":"2026-11-08T21:16:18.559Z"},"provenance":{"textHash":"ab0c3465e9a1e0bdaadbd5ecdc8875df8a0939f122deb944eeac559d50b66b53","observationId":"170000c695d08d634409fa270f1ec8aab2434d9846801875e124cbd4d8ea5471","retrievedAt":"2026-11-08T21:16:17.568Z","kind":"publisher-page","preparer":{"model":"gpt-6-luna","runId":"36a8a050-fdec-471f-8212-50c051230930"},"reviewer":{"model":"gpt-6-luna","runId":"f473c3ce-6ba5-4653-8642-fd3ce45c36a5"}},"audit":{"status":"uncertain","reviewedAt":"2026-11-08T21:17:05.469Z","model":"anthropic/claude-opus-5.5"},"sourceAttribution":{"authors":[{"name":"Shi, Yunxiao","url":null,"kind":"unknown"},{"name":"Xu, Wujiang","url":null,"kind":"unknown"},{"name":"Chen, Tingwei","url":null,"kind":"unknown"},{"name":"Shang, Haoning","url":null,"kind":"unknown"},{"name":"Yang, Ling","url":null,"kind":"unknown"},{"name":"Wan, Yunfeng","url":null,"kind":"unknown"},{"name":"Cao, Zhuo","url":null,"kind":"unknown"},{"name":"Zi, Xing","url":null,"kind":"unknown"},{"name":"Metaxas, Dimitris N.","url":null,"kind":"unknown"},{"name":"Xu, Min","url":null,"kind":"unknown"}],"authorRole":"article-author","hosts":[],"status":"stated","authorRoster":{"truncated":false,"omitted":0,"assertionGroupsOmitted":0},"participation":[{"actor":{"name":"Shi, Yunxiao","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Shi, Yunxiao","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2603.03761","locator":"head/meta[name=author|citation_author]"},"assertionId":"f11c3ada3041d31719c30c203e2d809ecaf7692749e6511c4699584e5c6c8d3d:participation:0","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-08T21:48:07.989Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"f11c3ada3041d31719c30c203e2d809ecaf7692749e6511c4699584e5c6c8d3d","sourceUrl":"https://arxiv.org/abs/2603.03761","contentSha256":"07395af2ff8798b35c8b7f5ab3f4d462058822cd14ee4e7d8e6340f236e6b673","textHash":"ab0c3465e9a1e0bdaadbd5ecdc8875df8a0939f122deb944eeac559d50b66b53","observedAt":"2026-11-08T21:48:07.989Z"}},{"actor":{"name":"Xu, Wujiang","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Xu, Wujiang","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2603.03761","locator":"head/meta[name=author|citation_author]"},"assertionId":"f11c3ada3041d31719c30c203e2d809ecaf7692749e6511c4699584e5c6c8d3d:participation:1","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-08T21:48:07.989Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"f11c3ada3041d31719c30c203e2d809ecaf7692749e6511c4699584e5c6c8d3d","sourceUrl":"https://arxiv.org/abs/2603.03761","contentSha256":"07395af2ff8798b35c8b7f5ab3f4d462058822cd14ee4e7d8e6340f236e6b673","textHash":"ab0c3465e9a1e0bdaadbd5ecdc8875df8a0939f122deb944eeac559d50b66b53","observedAt":"2026-11-08T21:48:07.989Z"}},{"actor":{"name":"Chen, Tingwei","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Chen, Tingwei","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2603.03761","locator":"head/meta[name=author|citation_author]"},"assertionId":"f11c3ada3041d31719c30c203e2d809ecaf7692749e6511c4699584e5c6c8d3d:participation:2","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-08T21:48:07.989Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"f11c3ada3041d31719c30c203e2d809ecaf7692749e6511c4699584e5c6c8d3d","sourceUrl":"https://arxiv.org/abs/2603.03761","contentSha256":"07395af2ff8798b35c8b7f5ab3f4d462058822cd14ee4e7d8e6340f236e6b673","textHash":"ab0c3465e9a1e0bdaadbd5ecdc8875df8a0939f122deb944eeac559d50b66b53","observedAt":"2026-11-08T21:48:07.989Z"}},{"actor":{"name":"Shang, Haoning","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Shang, Haoning","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2603.03761","locator":"head/meta[name=author|citation_author]"},"assertionId":"f11c3ada3041d31719c30c203e2d809ecaf7692749e6511c4699584e5c6c8d3d:participation:3","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-08T21:48:07.989Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"f11c3ada3041d31719c30c203e2d809ecaf7692749e6511c4699584e5c6c8d3d","sourceUrl":"https://arxiv.org/abs/2603.03761","contentSha256":"07395af2ff8798b35c8b7f5ab3f4d462058822cd14ee4e7d8e6340f236e6b673","textHash":"ab0c3465e9a1e0bdaadbd5ecdc8875df8a0939f122deb944eeac559d50b66b53","observedAt":"2026-11-08T21:48:07.989Z"}},{"actor":{"name":"Yang, Ling","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Yang, Ling","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2603.03761","locator":"head/meta[name=author|citation_author]"},"assertionId":"f11c3ada3041d31719c30c203e2d809ecaf7692749e6511c4699584e5c6c8d3d:participation:4","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-08T21:48:07.989Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"f11c3ada3041d31719c30c203e2d809ecaf7692749e6511c4699584e5c6c8d3d","sourceUrl":"https://arxiv.org/abs/2603.03761","contentSha256":"07395af2ff8798b35c8b7f5ab3f4d462058822cd14ee4e7d8e6340f236e6b673","textHash":"ab0c3465e9a1e0bdaadbd5ecdc8875df8a0939f122deb944eeac559d50b66b53","observedAt":"2026-11-08T21:48:07.989Z"}},{"actor":{"name":"Wan, Yunfeng","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Wan, Yunfeng","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2603.03761","locator":"head/meta[name=author|citation_author]"},"assertionId":"f11c3ada3041d31719c30c203e2d809ecaf7692749e6511c4699584e5c6c8d3d:participation:5","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-08T21:48:07.989Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"f11c3ada3041d31719c30c203e2d809ecaf7692749e6511c4699584e5c6c8d3d","sourceUrl":"https://arxiv.org/abs/2603.03761","contentSha256":"07395af2ff8798b35c8b7f5ab3f4d462058822cd14ee4e7d8e6340f236e6b673","textHash":"ab0c3465e9a1e0bdaadbd5ecdc8875df8a0939f122deb944eeac559d50b66b53","observedAt":"2026-11-08T21:48:07.989Z"}},{"actor":{"name":"Cao, Zhuo","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Cao, Zhuo","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2603.03761","locator":"head/meta[name=author|citation_author]"},"assertionId":"f11c3ada3041d31719c30c203e2d809ecaf7692749e6511c4699584e5c6c8d3d:participation:6","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-08T21:48:07.989Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"f11c3ada3041d31719c30c203e2d809ecaf7692749e6511c4699584e5c6c8d3d","sourceUrl":"https://arxiv.org/abs/2603.03761","contentSha256":"07395af2ff8798b35c8b7f5ab3f4d462058822cd14ee4e7d8e6340f236e6b673","textHash":"ab0c3465e9a1e0bdaadbd5ecdc8875df8a0939f122deb944eeac559d50b66b53","observedAt":"2026-11-08T21:48:07.989Z"}},{"actor":{"name":"Zi, Xing","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Zi, Xing","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2603.03761","locator":"head/meta[name=author|citation_author]"},"assertionId":"f11c3ada3041d31719c30c203e2d809ecaf7692749e6511c4699584e5c6c8d3d:participation:7","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-08T21:48:07.989Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"f11c3ada3041d31719c30c203e2d809ecaf7692749e6511c4699584e5c6c8d3d","sourceUrl":"https://arxiv.org/abs/2603.03761","contentSha256":"07395af2ff8798b35c8b7f5ab3f4d462058822cd14ee4e7d8e6340f236e6b673","textHash":"ab0c3465e9a1e0bdaadbd5ecdc8875df8a0939f122deb944eeac559d50b66b53","observedAt":"2026-11-08T21:48:07.989Z"}},{"actor":{"name":"Metaxas, Dimitris N.","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Metaxas, Dimitris N.","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2603.03761","locator":"head/meta[name=author|citation_author]"},"assertionId":"f11c3ada3041d31719c30c203e2d809ecaf7692749e6511c4699584e5c6c8d3d:participation:8","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-08T21:48:07.989Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"f11c3ada3041d31719c30c203e2d809ecaf7692749e6511c4699584e5c6c8d3d","sourceUrl":"https://arxiv.org/abs/2603.03761","contentSha256":"07395af2ff8798b35c8b7f5ab3f4d462058822cd14ee4e7d8e6340f236e6b673","textHash":"ab0c3465e9a1e0bdaadbd5ecdc8875df8a0939f122deb944eeac559d50b66b53","observedAt":"2026-11-08T21:48:07.989Z"}},{"actor":{"name":"Xu, Min","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Xu, Min","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2603.03761","locator":"head/meta[name=author|citation_author]"},"assertionId":"f11c3ada3041d31719c30c203e2d809ecaf7692749e6511c4699584e5c6c8d3d:participation:9","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-08T21:48:07.989Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"f11c3ada3041d31719c30c203e2d809ecaf7692749e6511c4699584e5c6c8d3d","sourceUrl":"https://arxiv.org/abs/2603.03761","contentSha256":"07395af2ff8798b35c8b7f5ab3f4d462058822cd14ee4e7d8e6340f236e6b673","textHash":"ab0c3465e9a1e0bdaadbd5ecdc8875df8a0939f122deb944eeac559d50b66b53","observedAt":"2026-11-08T21:48:07.989Z"}}],"participationCoverage":{"total":10,"returned":10,"complete":true,"fullAssertions":"source-metadata-endpoint"},"distribution":null,"publisher":null,"sourceUrl":"https://arxiv.org/abs/2603.03761","snapshotId":"f11c3ada3041d31719c30c203e2d809ecaf7692749e6511c4699584e5c6c8d3d","contentSha256":"07395af2ff8798b35c8b7f5ab3f4d462058822cd14ee4e7d8e6340f236e6b673","textHash":"ab0c3465e9a1e0bdaadbd5ecdc8875df8a0939f122deb944eeac559d50b66b53","observedAt":"2026-11-08T21:48:07.989Z","basis":"publisher-page-metadata","sourceLinks":51,"linksTruncated":false,"metadataUrl":"/api/pipeline/stream/4e1ab952f26a86686b8ddce532ab9865f4d41f63470c8050793a36a0aaea1e37/source"}}}],"audits":[{"publicationId":"9ac46481-dce2-4edc-8c03-27fe6d25d5cf","verdict":"uncertain","findings":[{"dimension":"attribution","status":"pass","explanation":"The reading attributes the problem framing, benchmark design and results to the AgentSelect authors through an arXiv paper, and keeps the findings as reported claims rather than established facts.","excerpt":"Authors:Yunxiao Shi, Wujiang Xu, Tingwei Chen, Haoning Shang, Ling Yang, Yunfeng Wan, Zhuo Cao, Xing Zi, Dimitris N. Metaxas, Min Xu"},{"dimension":"summary","status":"uncertain","explanation":"The summary is mostly faithful, but its central finding uses the jargon 'popularity-based CF/GNN methods' without explaining it for non-specialists. It also changes the source term 'content-aware capability matching' to 'capability-aware matching.' That loses the key contrast with popularity-based methods. Smallest repair: keep the exact source term and add tooltips for CF (collaborative filtering), GNN (graph neural network) and long-tail.","excerpt":"popularity-based CF/GNN methods become fragile and content-aware capability matching is essential"},{"dimension":"date","status":"pass","explanation":"originalPublished is 2026-03-04, which matches the v1 submission date. The 5 Nov 2026 date is a later revision, and the feed date is a later announcement.","excerpt":"[Submitted on 4 Mar 2026 (v1), last revised 5 Nov 2026 (this version, v2)]"},{"dimension":"support","status":"pass","explanation":"The dataset counts, the gap framing, the long-tail finding, the learnability of synthesized interactions and the MuleRun transfer are all stated in the abstract. The closing caveat and the question about how consistent the gains are do not overreach. They correctly mark what the abstract leaves unshown.","excerpt":"models trained on AgentSelect also transfer to a public agent marketplace (MuleRun), yielding consistent gains on an unseen catalog"},{"dimension":"media","status":"pass","explanation":"The selected media is the verified, direct arXiv abstract page for this paper, which is the most direct way to reach the source.","excerpt":null}],"createdAt":"2026-11-08T21:17:05.469Z","model":"anthropic/claude-opus-5.5"}],"correctionUrl":"https://escapemeridian.com/corrections/"}