{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:4ASF3VZCDZDCXY5SRSRPO66ZPG","short_pith_number":"pith:4ASF3VZC","schema_version":"1.0","canonical_sha256":"e0245dd7221e462be3b28ca2f77bd979baf384bd1654ef2a5a62415ae5802847","source":{"kind":"arxiv","id":"2607.15524","version":1},"attestation_state":"computed","paper":{"title":"Recursive Harness Self-Improvement","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Donghyun Lee, Hyunin Lee, Jeffrey Seely, Jinglue Xu, Matei Zaharia, Yujin Tang","submitted_at":"2026-07-17T00:21:19Z","abstract_excerpt":"Under model--harness co-evolution, harnesses are not merely inference-time scaffolds but data-generating components whose execution traces can shape future foundation models. This motivates harness-in-the-loop learning: optimizing harnesses for both immediate agent performance and the quality of traces used for future model training. However, continually updating provider-built scaffolds is costly and labor-intensive. We therefore investigate whether optimizing user-constructed harnesses in a task-specific manner can improve execution-trace quality while remaining computationally lightweight 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":"2607.15524","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-07-17T00:21:19Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"0b94c9fe86503a763b4a122658904fdb4cca75191e3fb14d46cb6df10fbfe66e","abstract_canon_sha256":"941437fb28049e405ca23967613137b77db33fc76810701c8baf1276f23bdbda"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-20T00:18:38.229832Z","signature_b64":"Bmsq/y78qYmFItjwFkchY3vPhYzqsDgq0rCRYIYPgu2QeYZC1N/gC7VcPgdaebAZSKyrsvpU62qt3ME4Jok1BA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e0245dd7221e462be3b28ca2f77bd979baf384bd1654ef2a5a62415ae5802847","last_reissued_at":"2026-07-20T00:18:38.228790Z","signature_status":"signed_v1","first_computed_at":"2026-07-20T00:18:38.228790Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Recursive Harness Self-Improvement","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Donghyun Lee, Hyunin Lee, Jeffrey Seely, Jinglue Xu, Matei Zaharia, Yujin Tang","submitted_at":"2026-07-17T00:21:19Z","abstract_excerpt":"Under model--harness co-evolution, harnesses are not merely inference-time scaffolds but data-generating components whose execution traces can shape future foundation models. This motivates harness-in-the-loop learning: optimizing harnesses for both immediate agent performance and the quality of traces used for future model training. However, continually updating provider-built scaffolds is costly and labor-intensive. We therefore investigate whether optimizing user-constructed harnesses in a task-specific manner can improve execution-trace quality while remaining computationally lightweight a"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2607.15524","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/2607.15524/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":"2607.15524","created_at":"2026-07-20T00:18:38.229264+00:00"},{"alias_kind":"arxiv_version","alias_value":"2607.15524v1","created_at":"2026-07-20T00:18:38.229264+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.15524","created_at":"2026-07-20T00:18:38.229264+00:00"},{"alias_kind":"pith_short_12","alias_value":"4ASF3VZCDZDC","created_at":"2026-07-20T00:18:38.229264+00:00"},{"alias_kind":"pith_short_16","alias_value":"4ASF3VZCDZDCXY5S","created_at":"2026-07-20T00:18:38.229264+00:00"},{"alias_kind":"pith_short_8","alias_value":"4ASF3VZC","created_at":"2026-07-20T00:18:38.229264+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/4ASF3VZCDZDCXY5SRSRPO66ZPG","json":"https://pith.science/pith/4ASF3VZCDZDCXY5SRSRPO66ZPG.json","graph_json":"https://pith.science/api/pith-number/4ASF3VZCDZDCXY5SRSRPO66ZPG/graph.json","events_json":"https://pith.science/api/pith-number/4ASF3VZCDZDCXY5SRSRPO66ZPG/events.json","paper":"https://pith.science/paper/4ASF3VZC"},"agent_actions":{"view_html":"https://pith.science/pith/4ASF3VZCDZDCXY5SRSRPO66ZPG","download_json":"https://pith.science/pith/4ASF3VZCDZDCXY5SRSRPO66ZPG.json","view_paper":"https://pith.science/paper/4ASF3VZC","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2607.15524&json=true","fetch_graph":"https://pith.science/api/pith-number/4ASF3VZCDZDCXY5SRSRPO66ZPG/graph.json","fetch_events":"https://pith.science/api/pith-number/4ASF3VZCDZDCXY5SRSRPO66ZPG/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/4ASF3VZCDZDCXY5SRSRPO66ZPG/action/timestamp_anchor","attest_storage":"https://pith.science/pith/4ASF3VZCDZDCXY5SRSRPO66ZPG/action/storage_attestation","attest_author":"https://pith.science/pith/4ASF3VZCDZDCXY5SRSRPO66ZPG/action/author_attestation","sign_citation":"https://pith.science/pith/4ASF3VZCDZDCXY5SRSRPO66ZPG/action/citation_signature","submit_replication":"https://pith.science/pith/4ASF3VZCDZDCXY5SRSRPO66ZPG/action/replication_record"}},"created_at":"2026-07-20T00:18:38.229264+00:00","updated_at":"2026-07-20T00:18:38.229264+00:00"}