{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:7LYMBPRDQNO2PWYKFQ4SCGWNUA","short_pith_number":"pith:7LYMBPRD","schema_version":"1.0","canonical_sha256":"faf0c0be23835da7db0a2c39211acda0268f75153499f6e5816873904c2544ec","source":{"kind":"arxiv","id":"2605.23772","version":1},"attestation_state":"computed","paper":{"title":"Agentic Proving for Program Verification","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":["cs.LO","cs.PL","cs.SE"],"primary_cat":"cs.AI","authors_text":"Akhil Arora, Alessandro Sosso, Bas Spitters","submitted_at":"2026-05-22T15:41:27Z","abstract_excerpt":"Agentic systems have recently emerged as state-of-the-art approaches for automated theorem proving in formal mathematics. To assess how far these capabilities extend to program verification, we evaluate Claude Code in an agentic proving framework on CLEVER, a Lean 4 benchmark for verifiable code generation. Our results show that Claude generates arguably valid specifications for 98.8% of problems (with 81.3% also accepted by CLEVER's isomorphism-based scoring on the correct portion of the benchmark), certifies implementations against correct ground-truth specifications for 87.5% of problems, a"},"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":"2605.23772","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.AI","submitted_at":"2026-05-22T15:41:27Z","cross_cats_sorted":["cs.LO","cs.PL","cs.SE"],"title_canon_sha256":"9ddf4b2eb1ae58b9f75e80e3137f3b1eb523d9ed4c2acad9839cf46f5973d285","abstract_canon_sha256":"34d7625842c9a725290db6fe3088a3ff4d32135c5759671e88799a783ceff5f6"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-25T02:02:31.811234Z","signature_b64":"irTp2uO5WbkDEMTzY/CTpawUXn77wIgVX8GXpldPjCZouNT57vopaT1iKRl50CwcWnB0151nBvrWJM7Vw9eMBA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"faf0c0be23835da7db0a2c39211acda0268f75153499f6e5816873904c2544ec","last_reissued_at":"2026-05-25T02:02:31.809773Z","signature_status":"signed_v1","first_computed_at":"2026-05-25T02:02:31.809773Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Agentic Proving for Program Verification","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":["cs.LO","cs.PL","cs.SE"],"primary_cat":"cs.AI","authors_text":"Akhil Arora, Alessandro Sosso, Bas Spitters","submitted_at":"2026-05-22T15:41:27Z","abstract_excerpt":"Agentic systems have recently emerged as state-of-the-art approaches for automated theorem proving in formal mathematics. To assess how far these capabilities extend to program verification, we evaluate Claude Code in an agentic proving framework on CLEVER, a Lean 4 benchmark for verifiable code generation. Our results show that Claude generates arguably valid specifications for 98.8% of problems (with 81.3% also accepted by CLEVER's isomorphism-based scoring on the correct portion of the benchmark), certifies implementations against correct ground-truth specifications for 87.5% of problems, a"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.23772","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/2605.23772/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":"2605.23772","created_at":"2026-05-25T02:02:31.810308+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.23772v1","created_at":"2026-05-25T02:02:31.810308+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.23772","created_at":"2026-05-25T02:02:31.810308+00:00"},{"alias_kind":"pith_short_12","alias_value":"7LYMBPRDQNO2","created_at":"2026-05-25T02:02:31.810308+00:00"},{"alias_kind":"pith_short_16","alias_value":"7LYMBPRDQNO2PWYK","created_at":"2026-05-25T02:02:31.810308+00:00"},{"alias_kind":"pith_short_8","alias_value":"7LYMBPRD","created_at":"2026-05-25T02:02:31.810308+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/7LYMBPRDQNO2PWYKFQ4SCGWNUA","json":"https://pith.science/pith/7LYMBPRDQNO2PWYKFQ4SCGWNUA.json","graph_json":"https://pith.science/api/pith-number/7LYMBPRDQNO2PWYKFQ4SCGWNUA/graph.json","events_json":"https://pith.science/api/pith-number/7LYMBPRDQNO2PWYKFQ4SCGWNUA/events.json","paper":"https://pith.science/paper/7LYMBPRD"},"agent_actions":{"view_html":"https://pith.science/pith/7LYMBPRDQNO2PWYKFQ4SCGWNUA","download_json":"https://pith.science/pith/7LYMBPRDQNO2PWYKFQ4SCGWNUA.json","view_paper":"https://pith.science/paper/7LYMBPRD","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.23772&json=true","fetch_graph":"https://pith.science/api/pith-number/7LYMBPRDQNO2PWYKFQ4SCGWNUA/graph.json","fetch_events":"https://pith.science/api/pith-number/7LYMBPRDQNO2PWYKFQ4SCGWNUA/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/7LYMBPRDQNO2PWYKFQ4SCGWNUA/action/timestamp_anchor","attest_storage":"https://pith.science/pith/7LYMBPRDQNO2PWYKFQ4SCGWNUA/action/storage_attestation","attest_author":"https://pith.science/pith/7LYMBPRDQNO2PWYKFQ4SCGWNUA/action/author_attestation","sign_citation":"https://pith.science/pith/7LYMBPRDQNO2PWYKFQ4SCGWNUA/action/citation_signature","submit_replication":"https://pith.science/pith/7LYMBPRDQNO2PWYKFQ4SCGWNUA/action/replication_record"}},"created_at":"2026-05-25T02:02:31.810308+00:00","updated_at":"2026-05-25T02:02:31.810308+00:00"}