{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:DD5JCR2ELDXB3R5DOQVXDHD3CU","short_pith_number":"pith:DD5JCR2E","schema_version":"1.0","canonical_sha256":"18fa91474458ee1dc7a3742b719c7b1538d040b2b3d1a6d8cc78f59cde35bd89","source":{"kind":"arxiv","id":"2506.00618","version":3},"attestation_state":"computed","paper":{"title":"RiOSWorld: Benchmarking the Risk of Multimodal Computer-Use Agents","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Dongrui Liu, Jing Shao, Jingyi Yang, Shuai Shao","submitted_at":"2025-05-31T16:04:59Z","abstract_excerpt":"With the rapid development of multimodal large language models (MLLMs), they are increasingly deployed as autonomous computer-use agents capable of accomplishing complex computer tasks. However, a pressing issue arises: Can the safety risk principles designed and aligned for general MLLMs in dialogue scenarios be effectively transferred to real-world computer-use scenarios? Existing research on evaluating the safety risks of MLLM-based computer-use agents suffers from several limitations: it either lacks realistic interactive environments, or narrowly focuses on one or a few specific risk type"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2506.00618","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2025-05-31T16:04:59Z","cross_cats_sorted":[],"title_canon_sha256":"cf081dc6ce1f11aff5476cb6a0f5c192797c8d430772536840fe0c267006b89f","abstract_canon_sha256":"fe894638e6e5d99c0f7fb117351f03625ddcad07f645deee13e9c4d880dcf60a"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:24:21.549033Z","signature_b64":"v+QYWMlpxUh2hLeFiaP1CM2ny2S58TO4JG23r7+3rWLrjD+C1b3qtNgHvoMR46DyHfjpaoNwXSIUuReKbFjJAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"18fa91474458ee1dc7a3742b719c7b1538d040b2b3d1a6d8cc78f59cde35bd89","last_reissued_at":"2026-07-05T11:24:21.548256Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:24:21.548256Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"RiOSWorld: Benchmarking the Risk of Multimodal Computer-Use Agents","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Dongrui Liu, Jing Shao, Jingyi Yang, Shuai Shao","submitted_at":"2025-05-31T16:04:59Z","abstract_excerpt":"With the rapid development of multimodal large language models (MLLMs), they are increasingly deployed as autonomous computer-use agents capable of accomplishing complex computer tasks. However, a pressing issue arises: Can the safety risk principles designed and aligned for general MLLMs in dialogue scenarios be effectively transferred to real-world computer-use scenarios? Existing research on evaluating the safety risks of MLLM-based computer-use agents suffers from several limitations: it either lacks realistic interactive environments, or narrowly focuses on one or a few specific risk type"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.00618","kind":"arxiv","version":3},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2506.00618/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"aliases":[{"alias_kind":"arxiv","alias_value":"2506.00618","created_at":"2026-07-05T11:24:21.548321+00:00"},{"alias_kind":"arxiv_version","alias_value":"2506.00618v3","created_at":"2026-07-05T11:24:21.548321+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.00618","created_at":"2026-07-05T11:24:21.548321+00:00"},{"alias_kind":"pith_short_12","alias_value":"DD5JCR2ELDXB","created_at":"2026-07-05T11:24:21.548321+00:00"},{"alias_kind":"pith_short_16","alias_value":"DD5JCR2ELDXB3R5D","created_at":"2026-07-05T11:24:21.548321+00:00"},{"alias_kind":"pith_short_8","alias_value":"DD5JCR2E","created_at":"2026-07-05T11:24:21.548321+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":6,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.31551","citing_title":"AutoTrainess: Teaching Language Models to Improve Language Models Autonomously","ref_index":18,"is_internal_anchor":false},{"citing_arxiv_id":"2605.13631","citing_title":"ProjGuard: Safety Monitoring for Computer-Use Agents via Low-Dimensional Projections","ref_index":33,"is_internal_anchor":false},{"citing_arxiv_id":"2606.29537","citing_title":"OSWorld 2.0: Benchmarking Computer Use Agents on Long-Horizon Real-World Tasks","ref_index":99,"is_internal_anchor":false},{"citing_arxiv_id":"2605.07110","citing_title":"Securing Computer-Use Agents: A Unified Architecture-Lifecycle Framework for Deployment-Grounded Reliability","ref_index":50,"is_internal_anchor":false},{"citing_arxiv_id":"2605.07630","citing_title":"Safe, or Simply Incapable? Rethinking Safety Evaluation for Phone-Use Agents","ref_index":15,"is_internal_anchor":false},{"citing_arxiv_id":"2604.13531","citing_title":"RiskWebWorld: A Realistic Interactive Benchmark for GUI Agents in E-commerce Risk Management","ref_index":56,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/DD5JCR2ELDXB3R5DOQVXDHD3CU","json":"https://pith.science/pith/DD5JCR2ELDXB3R5DOQVXDHD3CU.json","graph_json":"https://pith.science/api/pith-number/DD5JCR2ELDXB3R5DOQVXDHD3CU/graph.json","events_json":"https://pith.science/api/pith-number/DD5JCR2ELDXB3R5DOQVXDHD3CU/events.json","paper":"https://pith.science/paper/DD5JCR2E"},"agent_actions":{"view_html":"https://pith.science/pith/DD5JCR2ELDXB3R5DOQVXDHD3CU","download_json":"https://pith.science/pith/DD5JCR2ELDXB3R5DOQVXDHD3CU.json","view_paper":"https://pith.science/paper/DD5JCR2E","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2506.00618&json=true","fetch_graph":"https://pith.science/api/pith-number/DD5JCR2ELDXB3R5DOQVXDHD3CU/graph.json","fetch_events":"https://pith.science/api/pith-number/DD5JCR2ELDXB3R5DOQVXDHD3CU/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/DD5JCR2ELDXB3R5DOQVXDHD3CU/action/timestamp_anchor","attest_storage":"https://pith.science/pith/DD5JCR2ELDXB3R5DOQVXDHD3CU/action/storage_attestation","attest_author":"https://pith.science/pith/DD5JCR2ELDXB3R5DOQVXDHD3CU/action/author_attestation","sign_citation":"https://pith.science/pith/DD5JCR2ELDXB3R5DOQVXDHD3CU/action/citation_signature","submit_replication":"https://pith.science/pith/DD5JCR2ELDXB3R5DOQVXDHD3CU/action/replication_record"}},"created_at":"2026-07-05T11:24:21.548321+00:00","updated_at":"2026-07-05T11:24:21.548321+00:00"}