{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:4ITB7L5B7XH6ZV7RH3QUZSQYKL","short_pith_number":"pith:4ITB7L5B","canonical_record":{"source":{"id":"2606.11042","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.AI","submitted_at":"2026-06-09T16:10:16Z","cross_cats_sorted":[],"title_canon_sha256":"d3da7b711a4d363fec3254a472a99781399744792b5847b094079c05b7e0f6d1","abstract_canon_sha256":"39c9d197a2a0a097dc4b9f0228e820163a2ed8f86e53e6b078e050439f187c23"},"schema_version":"1.0"},"canonical_sha256":"e2261fafa1fdcfecd7f13ee14cca1852f9f588983ce4df0bdd887275d00ed2c8","source":{"kind":"arxiv","id":"2606.11042","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2606.11042","created_at":"2026-06-10T01:11:06Z"},{"alias_kind":"arxiv_version","alias_value":"2606.11042v1","created_at":"2026-06-10T01:11:06Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.11042","created_at":"2026-06-10T01:11:06Z"},{"alias_kind":"pith_short_12","alias_value":"4ITB7L5B7XH6","created_at":"2026-06-10T01:11:06Z"},{"alias_kind":"pith_short_16","alias_value":"4ITB7L5B7XH6ZV7R","created_at":"2026-06-10T01:11:06Z"},{"alias_kind":"pith_short_8","alias_value":"4ITB7L5B","created_at":"2026-06-10T01:11:06Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:4ITB7L5B7XH6ZV7RH3QUZSQYKL","target":"record","payload":{"canonical_record":{"source":{"id":"2606.11042","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.AI","submitted_at":"2026-06-09T16:10:16Z","cross_cats_sorted":[],"title_canon_sha256":"d3da7b711a4d363fec3254a472a99781399744792b5847b094079c05b7e0f6d1","abstract_canon_sha256":"39c9d197a2a0a097dc4b9f0228e820163a2ed8f86e53e6b078e050439f187c23"},"schema_version":"1.0"},"canonical_sha256":"e2261fafa1fdcfecd7f13ee14cca1852f9f588983ce4df0bdd887275d00ed2c8","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-10T01:11:06.536276Z","signature_b64":"0hFzs3IB8CxKn+9HA0lP2GizYxyWhoSCxz1qXQLKqlERy0ufaFZy45yXyio+vwHSB7dqydaWWoJO0SUw/RxPBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e2261fafa1fdcfecd7f13ee14cca1852f9f588983ce4df0bdd887275d00ed2c8","last_reissued_at":"2026-06-10T01:11:06.535376Z","signature_status":"signed_v1","first_computed_at":"2026-06-10T01:11:06.535376Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2606.11042","source_version":1,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-06-10T01:11:06Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"gC1zHUGL/DyF27hoE4bfP1tupwYDopTuVTOyCP44DgIo/gPEX0i7NMDcVsVkZ0/b+gYdn1lpOjNYLe8cmsMdAg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-11T11:23:07.570105Z"},"content_sha256":"99d579bc4a1d158aad7b206e8dbcddd3102f7edc92143e99def6c5fb0ef99340","schema_version":"1.0","event_id":"sha256:99d579bc4a1d158aad7b206e8dbcddd3102f7edc92143e99def6c5fb0ef99340"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:4ITB7L5B7XH6ZV7RH3QUZSQYKL","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Workflow-GYM: Towards Long-Horizon Evaluation of Computer-use Agentic tasks in Real-World Professional Fields","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Chaobo Xian, Chao He, Chenchen Zhang, Chi Wu, Chuqian Yu, Fangzhi Xu, Ge Zhang, Guanhong Chen, Haodong Duan, Haojun Wang, Haozhe Wang, Hongwan Gao, Huan Zhou, Huimin Che, Jiale Yang, Jiamin Chen, Jianbo Xue, Jiani Hou, Jian Zhang, Jiaying Wang, Jin Chen, Jingzhe Ding, Kaiyuan Zhang, Lili Le, Liya Zhu, Mailun Gao, Meng Cao, Minchao Wang, Mingzhang Wang, Peiheng Zhou, Pu Chen, Qing Hua, Qingshui Gu, Shengda Long, Shen Yan, Shihao Liang, Sijin Wu, Tianfeng Long, Tingqin Luo, Weiran Shi, Wenhao Huang, Xiang Gao, Xiaolong Chang, Xingyu Li, Yan Zhao, Yichen Zhang, Yiming Dai, Yiwei Liu, Yi Zhu, Yuhao Jiang, Yujia Qin, Zaiyuan Wang, Zhaojian Li, Zhengxuan Jiang, Zhenyu Hu, Zhixin Yao, Zhiyong Wu","submitted_at":"2026-06-09T16:10:16Z","abstract_excerpt":"Recent years have witnessed the rapid evolution of AI agents toward handling increasingly complex, real-world tasks. However, existing benchmarks rarely evaluate whether agents can operate graphical user interfaces to complete long-horizon, high-value professional workflows across diverse domains. Current GUI benchmarks still predominantly focus on general-purpose software, relatively simple applications, and short-horizon tasks, leaving it largely unknown whether modern agents can follow user instructions to autonomously operate domain-specific professional software and accomplish economicall"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.11042","kind":"arxiv","version":1},"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/2606.11042/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"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-06-10T01:11:06Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"zU/Uu5p9cDB/t9oVqleVSpTnOuIHKZz05v+1ypqWYgd8nO18yYY58sv2WBXVmIpN7ctWJoyj2s676/v2NvFeDg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-11T11:23:07.570512Z"},"content_sha256":"d1550fb6e1f61e4f851e2c26041416b2a628b93a4900bd58d44efaa46ea6ead9","schema_version":"1.0","event_id":"sha256:d1550fb6e1f61e4f851e2c26041416b2a628b93a4900bd58d44efaa46ea6ead9"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/4ITB7L5B7XH6ZV7RH3QUZSQYKL/bundle.json","state_url":"https://pith.science/pith/4ITB7L5B7XH6ZV7RH3QUZSQYKL/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/4ITB7L5B7XH6ZV7RH3QUZSQYKL/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-06-11T11:23:07Z","links":{"resolver":"https://pith.science/pith/4ITB7L5B7XH6ZV7RH3QUZSQYKL","bundle":"https://pith.science/pith/4ITB7L5B7XH6ZV7RH3QUZSQYKL/bundle.json","state":"https://pith.science/pith/4ITB7L5B7XH6ZV7RH3QUZSQYKL/state.json","well_known_bundle":"https://pith.science/.well-known/pith/4ITB7L5B7XH6ZV7RH3QUZSQYKL/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:4ITB7L5B7XH6ZV7RH3QUZSQYKL","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"39c9d197a2a0a097dc4b9f0228e820163a2ed8f86e53e6b078e050439f187c23","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.AI","submitted_at":"2026-06-09T16:10:16Z","title_canon_sha256":"d3da7b711a4d363fec3254a472a99781399744792b5847b094079c05b7e0f6d1"},"schema_version":"1.0","source":{"id":"2606.11042","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2606.11042","created_at":"2026-06-10T01:11:06Z"},{"alias_kind":"arxiv_version","alias_value":"2606.11042v1","created_at":"2026-06-10T01:11:06Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.11042","created_at":"2026-06-10T01:11:06Z"},{"alias_kind":"pith_short_12","alias_value":"4ITB7L5B7XH6","created_at":"2026-06-10T01:11:06Z"},{"alias_kind":"pith_short_16","alias_value":"4ITB7L5B7XH6ZV7R","created_at":"2026-06-10T01:11:06Z"},{"alias_kind":"pith_short_8","alias_value":"4ITB7L5B","created_at":"2026-06-10T01:11:06Z"}],"graph_snapshots":[{"event_id":"sha256:d1550fb6e1f61e4f851e2c26041416b2a628b93a4900bd58d44efaa46ea6ead9","target":"graph","created_at":"2026-06-10T01:11:06Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2606.11042/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Recent years have witnessed the rapid evolution of AI agents toward handling increasingly complex, real-world tasks. However, existing benchmarks rarely evaluate whether agents can operate graphical user interfaces to complete long-horizon, high-value professional workflows across diverse domains. Current GUI benchmarks still predominantly focus on general-purpose software, relatively simple applications, and short-horizon tasks, leaving it largely unknown whether modern agents can follow user instructions to autonomously operate domain-specific professional software and accomplish economicall","authors_text":"Chaobo Xian, Chao He, Chenchen Zhang, Chi Wu, Chuqian Yu, Fangzhi Xu, Ge Zhang, Guanhong Chen, Haodong Duan, Haojun Wang, Haozhe Wang, Hongwan Gao, Huan Zhou, Huimin Che, Jiale Yang, Jiamin Chen, Jianbo Xue, Jiani Hou, Jian Zhang, Jiaying Wang, Jin Chen, Jingzhe Ding, Kaiyuan Zhang, Lili Le, Liya Zhu, Mailun Gao, Meng Cao, Minchao Wang, Mingzhang Wang, Peiheng Zhou, Pu Chen, Qing Hua, Qingshui Gu, Shengda Long, Shen Yan, Shihao Liang, Sijin Wu, Tianfeng Long, Tingqin Luo, Weiran Shi, Wenhao Huang, Xiang Gao, Xiaolong Chang, Xingyu Li, Yan Zhao, Yichen Zhang, Yiming Dai, Yiwei Liu, Yi Zhu, Yuhao Jiang, Yujia Qin, Zaiyuan Wang, Zhaojian Li, Zhengxuan Jiang, Zhenyu Hu, Zhixin Yao, Zhiyong Wu","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.AI","submitted_at":"2026-06-09T16:10:16Z","title":"Workflow-GYM: Towards Long-Horizon Evaluation of Computer-use Agentic tasks in Real-World Professional Fields"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.11042","kind":"arxiv","version":1},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:99d579bc4a1d158aad7b206e8dbcddd3102f7edc92143e99def6c5fb0ef99340","target":"record","created_at":"2026-06-10T01:11:06Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"39c9d197a2a0a097dc4b9f0228e820163a2ed8f86e53e6b078e050439f187c23","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.AI","submitted_at":"2026-06-09T16:10:16Z","title_canon_sha256":"d3da7b711a4d363fec3254a472a99781399744792b5847b094079c05b7e0f6d1"},"schema_version":"1.0","source":{"id":"2606.11042","kind":"arxiv","version":1}},"canonical_sha256":"e2261fafa1fdcfecd7f13ee14cca1852f9f588983ce4df0bdd887275d00ed2c8","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"e2261fafa1fdcfecd7f13ee14cca1852f9f588983ce4df0bdd887275d00ed2c8","first_computed_at":"2026-06-10T01:11:06.535376Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-06-10T01:11:06.535376Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"0hFzs3IB8CxKn+9HA0lP2GizYxyWhoSCxz1qXQLKqlERy0ufaFZy45yXyio+vwHSB7dqydaWWoJO0SUw/RxPBg==","signature_status":"signed_v1","signed_at":"2026-06-10T01:11:06.536276Z","signed_message":"canonical_sha256_bytes"},"source_id":"2606.11042","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:99d579bc4a1d158aad7b206e8dbcddd3102f7edc92143e99def6c5fb0ef99340","sha256:d1550fb6e1f61e4f851e2c26041416b2a628b93a4900bd58d44efaa46ea6ead9"],"state_sha256":"c75506b4d3ee938799bebc043d11bb47dd660fc2a3439606e3d0a2bc0cf3735b"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"cNYqz0Z+QK5TQRul67uVYXDanhCGwblVJ4lHvKjKv1k5xJ77SZRsTPt29vCM2YAMHyES8wrJoHDxVLrr14RXDg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-06-11T11:23:07.572560Z","bundle_sha256":"727b0af58ad097b3375c7a18b88a58bfbaf6411c9dc39a2bc53ef2fa484bb0cf"}}