{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:FU3UE5RC7MRSMAJ472JIXZG2RS","short_pith_number":"pith:FU3UE5RC","schema_version":"1.0","canonical_sha256":"2d37427622fb2326013cfe928be4da8c8a690bc21815d29d6976a4c1a5fbb924","source":{"kind":"arxiv","id":"2606.22000","version":1},"attestation_state":"computed","paper":{"title":"CFAgentBench: A Reproducible Environment and Benchmark for Autonomous Construction-Finance Agents","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CL","cs.LG","cs.SE"],"primary_cat":"cs.AI","authors_text":"Rishi Srivastava","submitted_at":"2026-06-20T11:34:52Z","abstract_excerpt":"We introduce CFAgentBench, a reproducible, self-hostable environment and benchmark for autonomous construction-finance agents: a CFO/controller-class agent operating across the real software stack a US construction finance team runs - ERP, project management, email, documents, pay applications, payroll, certified payroll, lien waivers, and bank/treasury portals. It contains 1,014 machine-gradeable task specifications across 8 domains and 77 families, every family grounded in a real source; a self-validated subset of 40 tasks (54 with a project-management extension) is compiled into oracle-vali"},"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":"2606.22000","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2026-06-20T11:34:52Z","cross_cats_sorted":["cs.CL","cs.LG","cs.SE"],"title_canon_sha256":"c3cfda209845f01db7365e7e803a2f30cd893bdf46f59935e99c50f3d0622aa5","abstract_canon_sha256":"12b91124ac69bec4ae5973c0235b86c1f96d6f97d507a7b0c053884de7f707c7"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-23T02:13:05.293309Z","signature_b64":"5lIsFP0tbFnUhzJbbE20aQDNPPqc+/TuuakjGWkqA1RVj1pN7+CTdlKGO5G4nYwcGjw307Ng2mDJMuU54ng4Aw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"2d37427622fb2326013cfe928be4da8c8a690bc21815d29d6976a4c1a5fbb924","last_reissued_at":"2026-06-23T02:13:05.292879Z","signature_status":"signed_v1","first_computed_at":"2026-06-23T02:13:05.292879Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"CFAgentBench: A Reproducible Environment and Benchmark for Autonomous Construction-Finance Agents","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CL","cs.LG","cs.SE"],"primary_cat":"cs.AI","authors_text":"Rishi Srivastava","submitted_at":"2026-06-20T11:34:52Z","abstract_excerpt":"We introduce CFAgentBench, a reproducible, self-hostable environment and benchmark for autonomous construction-finance agents: a CFO/controller-class agent operating across the real software stack a US construction finance team runs - ERP, project management, email, documents, pay applications, payroll, certified payroll, lien waivers, and bank/treasury portals. It contains 1,014 machine-gradeable task specifications across 8 domains and 77 families, every family grounded in a real source; a self-validated subset of 40 tasks (54 with a project-management extension) is compiled into oracle-vali"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.22000","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/2606.22000/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":"2606.22000","created_at":"2026-06-23T02:13:05.292936+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.22000v1","created_at":"2026-06-23T02:13:05.292936+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.22000","created_at":"2026-06-23T02:13:05.292936+00:00"},{"alias_kind":"pith_short_12","alias_value":"FU3UE5RC7MRS","created_at":"2026-06-23T02:13:05.292936+00:00"},{"alias_kind":"pith_short_16","alias_value":"FU3UE5RC7MRSMAJ4","created_at":"2026-06-23T02:13:05.292936+00:00"},{"alias_kind":"pith_short_8","alias_value":"FU3UE5RC","created_at":"2026-06-23T02:13:05.292936+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/FU3UE5RC7MRSMAJ472JIXZG2RS","json":"https://pith.science/pith/FU3UE5RC7MRSMAJ472JIXZG2RS.json","graph_json":"https://pith.science/api/pith-number/FU3UE5RC7MRSMAJ472JIXZG2RS/graph.json","events_json":"https://pith.science/api/pith-number/FU3UE5RC7MRSMAJ472JIXZG2RS/events.json","paper":"https://pith.science/paper/FU3UE5RC"},"agent_actions":{"view_html":"https://pith.science/pith/FU3UE5RC7MRSMAJ472JIXZG2RS","download_json":"https://pith.science/pith/FU3UE5RC7MRSMAJ472JIXZG2RS.json","view_paper":"https://pith.science/paper/FU3UE5RC","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.22000&json=true","fetch_graph":"https://pith.science/api/pith-number/FU3UE5RC7MRSMAJ472JIXZG2RS/graph.json","fetch_events":"https://pith.science/api/pith-number/FU3UE5RC7MRSMAJ472JIXZG2RS/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/FU3UE5RC7MRSMAJ472JIXZG2RS/action/timestamp_anchor","attest_storage":"https://pith.science/pith/FU3UE5RC7MRSMAJ472JIXZG2RS/action/storage_attestation","attest_author":"https://pith.science/pith/FU3UE5RC7MRSMAJ472JIXZG2RS/action/author_attestation","sign_citation":"https://pith.science/pith/FU3UE5RC7MRSMAJ472JIXZG2RS/action/citation_signature","submit_replication":"https://pith.science/pith/FU3UE5RC7MRSMAJ472JIXZG2RS/action/replication_record"}},"created_at":"2026-06-23T02:13:05.292936+00:00","updated_at":"2026-06-23T02:13:05.292936+00:00"}