{"record": {"aliases": [], "artifacts": [{"id": "arxiv:2606.00925", "kind": "paper", "url": "https://arxiv.org/abs/2606.00925"}], "authors": ["Ismail Hossain", "Sai Puppala", "Zhuoran Lu", "Sajedul Talukder", "Nan Jiang"], "authors_reference": {"basis": "arxiv_metadata", "retrieved_at": "2026-11-03", "source_key": "claire-radar:2606.00925", "source_url": "https://arxiv.org/abs/2606.00925"}, "categories": ["General AI", "Vision & 3D", "cs.CR"], "description": {"en": "SkillVetBench is a two-stage security vetting benchmark for open agentic skill ecosystems, evaluating detection of malicious skills via semantic analysis and runtime verification in a sandbox. The benchmark is built from confirmed malicious skills in the OpenClaw ecosystem."}, "documents": [{"document_type": "registry_page", "id": "claire_radar:https://arxiv.org/abs/2606.00925", "source": "claire_radar", "source_url": "https://arxiv.org/abs/2606.00925", "title": "SkillVetBench"}], "evidence_summary": {"document_count": 1, "model_count": null, "model_count_basis": null}, "institutions": ["The University of Texas at El Paso", "Southern Illinois University Carbondale", "Purdue University West Lafayette"], "institutions_reference": {"basis": "openalex_linked_institutions", "doi": "10.48550/arxiv.2606.00925", "retrieved_at": "2026-11-04", "source_key": "claire-radar:2606.00925", "source_url": "https://openalex.org/W7163205394"}, "key": "claire-radar:2606.00925", "modality": null, "name": "SkillVetBench", "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_skillvetbench_809c89ae", "origin_source": "arxiv", "recognition_confidence": "1.0", "record_sha256": "8ee1ad993eedf00525966c67f4e0be72cac24907c8b26d9ad0517d5690c4a8a8", "relation": "introduces", "review_state": "unreviewed", "source_url": "https://arxiv.org/abs/2606.00925"}, "publisher": null, "released": "2026-05-30", "released_reference": {"basis": "benchmark_release", "source_key": "claire-radar:2606.00925", "source_url": "https://arxiv.org/abs/2606.00925"}, "schema_version": 1, "sizes": [], "slug": "claire-radar-2606-00925", "source": "claire_radar", "source_benchmark_id": "2606.00925", "source_metadata": {"claire_radar": {"adoption30d": null, "aliases": [], "annotation": "Unknown", "applicationDomains": ["General AI"], "area": "Vision & 3D", "attention": {"asOf": "2026-10-05", "attemptedAt": "2026-10-04T15:31:48.030916Z", "githubRepo": null, "githubScope": null, "githubStars": null, "hfDailySubmittedAt": null, "hfDataset": null, "hfDatasetDownloads": null, "hfDatasetLikes": null, "hfPaperUpvotes": null, "hfPaperUrl": "https://huggingface.co/papers/2606.00925", "observedAt": "2026-10-04T03:02:42.754940Z", "signalStatus": {"githubStars": {"state": "not_applicable"}, "hfDatasetDownloads": {"state": "not_applicable"}, "hfDatasetLikes": {"state": "not_applicable"}, "hfPaperUpvotes": {"checkedAt": "2026-10-04T15:31:48.030916Z", "state": "unavailable"}}, "status": "unavailable"}, "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": "6cd31b8641e63a6bd86c0c4d989c9c886dbfffc6108291f43641b0a1800463bb", "model": "deepseek-v4-flash"}, "dataStatus": "primary-source-indexed", "demo": false, "description": "SkillVetBench is a two-stage security vetting benchmark for open agentic skill ecosystems, evaluating detection of malicious skills via semantic analysis and runtime verification in a sandbox. The benchmark is built from confirmed malicious skills in the OpenClaw ecosystem.", "domainCuration": {"method": "rules-v1", "state": "auto"}, "evidence": {"reasonCodes": ["exact named benchmark artifact released in abstract", "benchmark release stated in one sentence", "evaluation protocol evidence"], "snippet": "We present SkillVetBench, a two-stage security vetting benchmark for open agentic skill ecosystems."}, "familyId": "bmf_13b32737c884", "firstSeenAt": "2026-08-19", "githubIndex": {"candidateCount": 0, "indexedAt": "2026-09-17T04:40:34.111665+00:00", "repositories": [], "status": "not-found"}, "heat": null, "id": "bm_skillvetbench_809c89ae", "indexedAt": "2026-08-19T11:05:13.395391Z", "industrySectors": [], "links": {"code": null, "data": null, "hfPaper": "https://huggingface.co/papers/2606.00925", "paper": "https://arxiv.org/abs/2606.00925", "pdf": "https://arxiv.org/pdf/2606.00925", "project": null, "report": "https://arxiv.org/abs/2606.00925"}, "metrics": [], "motivation": "Open agent platforms allow community contributors to publish reusable skills that agents can invoke at runtime.", "name": "SkillVetBench", "oneLine": "SkillVetBench is a two-stage security vetting benchmark for open agentic skill ecosystems, evaluating detection of malicious skills via semantic analysis and runtime verification in a sandbox. The benchmark is built from confirmed malicious skills in the OpenClaw ecosystem.", "paperTitle": "Benchmarking Security Risk Detection and Verification in Open Agentic Skill Ecosystems", "primaryDomain": "General AI", "publication": {"evidence": null, "evidenceLevel": "unverified", "evidenceUrl": "https://arxiv.org/abs/2606.00925", "source": "arxiv-metadata", "status": "unverified", "venue": null, "verifiedAt": "2026-10-14T01:46:58.058199Z"}, "publications": [], "ranking": {}, "readiness": "Paper only", "recognitionConfidence": 1.0, "relation": "introduces", "releasedAt": "2026-05-30", "source": {"authors": ["Ismail Hossain", "Sai Puppala", "Zhuoran Lu", "Sajedul Talukder and Nan Jiang"], "categories": ["cs.CR", "cs.AI"], "id": "2606.00925", "title": "Benchmarking Security Risk Detection and Verification in Open Agentic Skill Ecosystems", "type": "arxiv", "url": "https://arxiv.org/abs/2606.00925"}, "sourceUpdatedAt": "2026-05-30T23:19:30Z", "topics": ["cs.CR"], "venueAttempts": [], "whyItMatters": "Open agent platforms face supply-chain risks from malicious skills, but existing defenses lack a standardized evaluation. This benchmark addresses the gap in measuring both detection and runtime verification, offering practical value for improving security in agent ecosystems."}}}, "schema_version": 1, "scores_by_source": {}, "siblings": []}
