{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:SMZW4L6HRSHOV4HAMCWF5ELZ3N","short_pith_number":"pith:SMZW4L6H","schema_version":"1.0","canonical_sha256":"93336e2fc78c8eeaf0e060ac5e9179db50f031b173d278c335f5c47e89db3452","source":{"kind":"arxiv","id":"2607.19865","version":1},"attestation_state":"computed","paper":{"title":"DocOps: A Verifiable Benchmark for Autonomous Agents in Complex Document Operations","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CL","cs.LG"],"primary_cat":"cs.AI","authors_text":"Boxi Cao, Dawei Yin, Hongyu Lin, Jiazhen Jiang, Le Sun, Lingyong Yan, Shuaiqiang Wang, Xianpei Han, Yaojie Lu","submitted_at":"2026-07-22T07:52:59Z","abstract_excerpt":"As autonomous agents rapidly evolve, their ability to reliably manipulate ubiquitous digital documents has become critical for enabling general-purpose AI assistants and automating complex workspace workflows. In this paper, we introduce DocOps, a deterministically verifiable evaluation framework underpinned by a hierarchical taxonomy that deconstructs document operations inspired by real-world practices into atomic dimensions and escalating workflow complexities. Based on DocOps, we systematically evaluate representative closed- and open-source models across various agentic harnesses, reveali"},"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":"2607.19865","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2026-07-22T07:52:59Z","cross_cats_sorted":["cs.CL","cs.LG"],"title_canon_sha256":"3664791ae357fd5c2254cd2baaccf5ddbc08a85975980df47accf28c0b2150ac","abstract_canon_sha256":"a4fdeb256aa892bc1410802edff5e36b19d8e3871df60a342f9c570c3a7ae95a"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-23T01:24:04.551550Z","signature_b64":"T77SxrXbtQ4AvOYGjYsNlzO18gTSfSkXjPGLK6wLv1rm5P9KKCoxgrL554sJJQPvqc8IOnTZpeKDGIlrob1dBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"93336e2fc78c8eeaf0e060ac5e9179db50f031b173d278c335f5c47e89db3452","last_reissued_at":"2026-07-23T01:24:04.550713Z","signature_status":"signed_v1","first_computed_at":"2026-07-23T01:24:04.550713Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"DocOps: A Verifiable Benchmark for Autonomous Agents in Complex Document Operations","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CL","cs.LG"],"primary_cat":"cs.AI","authors_text":"Boxi Cao, Dawei Yin, Hongyu Lin, Jiazhen Jiang, Le Sun, Lingyong Yan, Shuaiqiang Wang, Xianpei Han, Yaojie Lu","submitted_at":"2026-07-22T07:52:59Z","abstract_excerpt":"As autonomous agents rapidly evolve, their ability to reliably manipulate ubiquitous digital documents has become critical for enabling general-purpose AI assistants and automating complex workspace workflows. In this paper, we introduce DocOps, a deterministically verifiable evaluation framework underpinned by a hierarchical taxonomy that deconstructs document operations inspired by real-world practices into atomic dimensions and escalating workflow complexities. Based on DocOps, we systematically evaluate representative closed- and open-source models across various agentic harnesses, reveali"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2607.19865","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/2607.19865/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":"2607.19865","created_at":"2026-07-23T01:24:04.551137+00:00"},{"alias_kind":"arxiv_version","alias_value":"2607.19865v1","created_at":"2026-07-23T01:24:04.551137+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.19865","created_at":"2026-07-23T01:24:04.551137+00:00"},{"alias_kind":"pith_short_12","alias_value":"SMZW4L6HRSHO","created_at":"2026-07-23T01:24:04.551137+00:00"},{"alias_kind":"pith_short_16","alias_value":"SMZW4L6HRSHOV4HA","created_at":"2026-07-23T01:24:04.551137+00:00"},{"alias_kind":"pith_short_8","alias_value":"SMZW4L6H","created_at":"2026-07-23T01:24:04.551137+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/SMZW4L6HRSHOV4HAMCWF5ELZ3N","json":"https://pith.science/pith/SMZW4L6HRSHOV4HAMCWF5ELZ3N.json","graph_json":"https://pith.science/api/pith-number/SMZW4L6HRSHOV4HAMCWF5ELZ3N/graph.json","events_json":"https://pith.science/api/pith-number/SMZW4L6HRSHOV4HAMCWF5ELZ3N/events.json","paper":"https://pith.science/paper/SMZW4L6H"},"agent_actions":{"view_html":"https://pith.science/pith/SMZW4L6HRSHOV4HAMCWF5ELZ3N","download_json":"https://pith.science/pith/SMZW4L6HRSHOV4HAMCWF5ELZ3N.json","view_paper":"https://pith.science/paper/SMZW4L6H","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2607.19865&json=true","fetch_graph":"https://pith.science/api/pith-number/SMZW4L6HRSHOV4HAMCWF5ELZ3N/graph.json","fetch_events":"https://pith.science/api/pith-number/SMZW4L6HRSHOV4HAMCWF5ELZ3N/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/SMZW4L6HRSHOV4HAMCWF5ELZ3N/action/timestamp_anchor","attest_storage":"https://pith.science/pith/SMZW4L6HRSHOV4HAMCWF5ELZ3N/action/storage_attestation","attest_author":"https://pith.science/pith/SMZW4L6HRSHOV4HAMCWF5ELZ3N/action/author_attestation","sign_citation":"https://pith.science/pith/SMZW4L6HRSHOV4HAMCWF5ELZ3N/action/citation_signature","submit_replication":"https://pith.science/pith/SMZW4L6HRSHOV4HAMCWF5ELZ3N/action/replication_record"}},"created_at":"2026-07-23T01:24:04.551137+00:00","updated_at":"2026-07-23T01:24:04.551137+00:00"}