{"record": {"aliases": [], "artifacts": [{"id": "arxiv:2608.06301", "kind": "paper", "url": "https://arxiv.org/abs/2608.06301"}], "authors": ["Varun Ursekar", "Apaar Shanker", "Yash Maurya", "Shehab Yasser", "Vijay S. Kalmath", "Veronica Chatrath", "Yuan Xue"], "authors_reference": {"basis": "arxiv_metadata", "retrieved_at": "2026-11-03", "source_key": "claire-radar:2608.06301", "source_url": "https://arxiv.org/abs/2608.06301"}, "categories": ["General AI", "Language & Knowledge", "cs.AI"], "description": {"en": "Evaluates LLMs optimizing a target agent's harness (prompts, tools, control flow) under budgeted, stochastic evaluation. Scoring is normalized gain over seed on a held-out test partition, with trusted execution environment enforcing evaluation boundary."}, "documents": [{"document_type": "registry_page", "id": "claire_radar:https://arxiv.org/abs/2608.06301", "source": "claire_radar", "source_url": "https://arxiv.org/abs/2608.06301", "title": "HarnessOpt-Bench"}], "evidence_summary": {"document_count": 1, "model_count": null, "model_count_basis": null}, "key": "claire-radar:2608.06301", "modality": null, "name": "HarnessOpt-Bench", "openness": {"code_license": null, "data_license": null, "evidence": [], "status": "unknown"}, "provenance": {"confidence": "High", "crawl_bundle": "claire_radar_2026-10-14", "crawled_at": "2026-10-14T09:44:51Z", "data_status": "primary-source-indexed", "display_eligible": "unknown", "origin_record_id": "bm_harnessopt-bench_24067463", "origin_source": "arxiv", "recognition_confidence": "1.0", "record_sha256": "73bc1c647fca4db663fb8cf34e3506432845388f62ea43c0d265d8a8d8d5f1c5", "relation": "introduces", "review_state": "unreviewed", "source_url": "https://arxiv.org/abs/2608.06301"}, "publisher": null, "released": "2026-08-06", "released_reference": {"basis": "benchmark_release", "source_key": "claire-radar:2608.06301", "source_url": "https://arxiv.org/abs/2608.06301"}, "schema_version": 1, "sizes": [], "slug": "claire-radar-2608-06301", "source": "claire_radar", "source_benchmark_id": "2608.06301", "source_metadata": {"claire_radar": {"adoption30d": null, "aliases": [], "annotation": "Unknown", "applicationDomains": ["General AI"], "area": "Language & Knowledge", "attention": {"asOf": "2026-10-05", "attemptedAt": "2026-10-04T15:31:48.030916Z", "githubRepo": null, "githubScope": null, "githubStars": null, "hfDailySubmittedAt": "2026-08-07T00:00:00.000Z", "hfDataset": null, "hfDatasetDownloads": null, "hfDatasetLikes": null, "hfPaperUpvotes": 35, "hfPaperUrl": "https://huggingface.co/papers/2608.06301", "observedAt": "2026-10-04T15:31:48.030916Z", "signalStatus": {"githubStars": {"state": "not_applicable"}, "hfDatasetDownloads": {"state": "not_applicable"}, "hfDatasetLikes": {"state": "not_applicable"}, "hfPaperUpvotes": {"observedAt": "2026-10-04T15:31:48.030916Z", "state": "fresh"}}, "status": "fresh"}, "capabilities": [], "confidence": "High", "construction": "Unknown", "constructionDetail": "Unknown — the indexer does not infer construction details without explicit source evidence.", "copyGeneration": {"generatedAt": "2026-08-20T15:22:10.044731Z", "inputHash": "cb38c3cb1edd81673d8340dd632183705f15e21f4f19da9f33d5572fb4765412", "model": "deepseek-v4-flash"}, "dataStatus": "primary-source-indexed", "demo": false, "description": "Evaluates LLMs optimizing a target agent's harness (prompts, tools, control flow) under budgeted, stochastic evaluation. Scoring is normalized gain over seed on a held-out test partition, with trusted execution environment enforcing evaluation boundary.", "domainCuration": {"method": "rules-v1", "state": "auto"}, "evidence": {"reasonCodes": ["coined title prefix ending in Bench or Benchmark", "benchmark release stated in one sentence", "evaluation protocol evidence"], "snippet": "We introduce HarnessOpt-Bench, a benchmark for end-to-end harness optimization under expensive and stochastic evaluation."}, "familyId": "bmf_42c79c547b84", "firstSeenAt": "2026-08-19", "githubIndex": {"candidateCount": 0, "indexedAt": "2026-09-17T04:40:34.111665+00:00", "repositories": [], "status": "not-found"}, "heat": null, "id": "bm_harnessopt-bench_24067463", "indexedAt": "2026-08-19T11:05:13.395391Z", "industrySectors": [], "links": {"code": null, "data": null, "hfPaper": "https://huggingface.co/papers/2608.06301", "paper": "https://arxiv.org/abs/2608.06301", "pdf": "https://arxiv.org/pdf/2608.06301", "project": null, "report": "https://arxiv.org/abs/2608.06301"}, "metrics": [], "motivation": "As LLMs are increasingly deployed within agentic systems, their capabilities depend not only on the model weights but also on the harness: the prompts, tools, control flow, memory, and orchestration code surrounding them.", "name": "HarnessOpt-Bench", "oneLine": "Evaluates LLMs optimizing a target agent's harness (prompts, tools, control flow) under budgeted, stochastic evaluation. Scoring is normalized gain over seed on a held-out test partition, with trusted execution environment enforcing evaluation boundary.", "paperTitle": "HarnessOpt-Bench: Evaluating LLMs at Harness Optimization", "primaryDomain": "General AI", "publication": {"evidence": null, "evidenceLevel": "unverified", "evidenceUrl": "https://arxiv.org/abs/2608.06301", "source": "arxiv-metadata", "status": "unverified", "venue": null, "verifiedAt": "2026-10-14T01:46:58.058199Z"}, "publications": [], "ranking": {"90d": {"confidence": "Low", "coverage": 0.15, "growth": {"components": {"githubStars": {"normalization": "percentile", "normalized": null, "percentile": null, "value": null}, "hfDatasetDownloads": {"normalization": "percentile", "normalized": null, "percentile": null, "value": null}, "hfPaperUpvotes": {"normalization": "percentile", "normalized": null, "percentile": null, "value": null}}, "confidence": "Low", "coverage": 0.0, "elapsedDays": null, "historyDate": null, "rank": null, "score": null, "windowDays": 90}, "level": {"components": {"githubStars": {"normalization": "log-max", "normalized": null, "value": null}, "hfDatasetDownloads": {"normalization": "log-max", "normalized": null, "value": null}, "hfPaperUpvotes": {"normalization": "log-max", "normalized": 0.6331884794120033, "value": 35}}, "confidence": "Low", "coverage": 0.15, "rank": null, "score": 52}, "method": "attention-ranking-v13", "rank": null, "score": 52}}, "readiness": "Paper only", "recognitionConfidence": 1.0, "relation": "introduces", "releasedAt": "2026-08-06", "source": {"authors": ["Varun Ursekar", "Apaar Shanker", "Yash Maurya", "Shehab Yasser", "Vijay S. Kalmath", "Veronica Chatrath", "Yuan Xue"], "categories": ["cs.AI", "cs.CL", "cs.LG"], "id": "2608.06301", "title": "HarnessOpt-Bench: Evaluating LLMs at Harness Optimization", "type": "arxiv", "url": "https://arxiv.org/abs/2608.06301"}, "sourceUpdatedAt": "2026-08-06T17:21:05Z", "topics": ["cs.AI"], "venueAttempts": [], "whyItMatters": "Addresses the gap in measuring automated harness optimization, a discriminative capability needed for improving agentic LLM systems. Provides a protocol for comparing optimizer models under varying harnesses and tasks."}}}, "schema_version": 1, "scores_by_source": {}, "siblings": []}
