{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:SGXTMIW2OCWCSUE3OAVK7ZLRDR","short_pith_number":"pith:SGXTMIW2","schema_version":"1.0","canonical_sha256":"91af3622da70ac29509b702aafe5711c773c327b636ede737c5e01889049f3ba","source":{"kind":"arxiv","id":"2607.02854","version":1},"attestation_state":"computed","paper":{"title":"EvoOtter: Evolutionary Reproduction Test Generator","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.SE","authors_text":"Avraham Shinnar, Jatin Ganhotra, Martin Hirzel, Toufique Ahmed","submitted_at":"2026-07-03T01:30:37Z","abstract_excerpt":"Before fixing an issue, it is useful to first reproduce it by generating a bug reproduction test (BRT). However, generating a BRT is itself a challenging task, because issue descriptions tend to be informal, making it difficult to determine whether a candidate BRT indeed fails for the reason in the issue. Prior work has attempted to tackle this problem via inference scaling, using large language models to generate many BRTs and patches, then using execution feedback to select and improve them. Unfortunately, this is expensive and the feedback is unreliable. This paper explores evolutionary pro"},"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":"2607.02854","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.SE","submitted_at":"2026-07-03T01:30:37Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"f6612547b2914f5ca14e0302f1e156701bb461ac6ce900e21c41cff8b6a5dd64","abstract_canon_sha256":"7f471283058b5b24dbf9dfe02cbd39dd6845893f00404bc3fd7a8bcf5204058e"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-07T00:16:20.844298Z","signature_b64":"Ls6BofSrUayNuP8CTOqG9tYr5AmM4OaimhwijqtmfkWYSEd7Zd+8yCWDurTktOdF48OW7IXOn33PAGuBzfokAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"91af3622da70ac29509b702aafe5711c773c327b636ede737c5e01889049f3ba","last_reissued_at":"2026-07-07T00:16:20.843646Z","signature_status":"signed_v1","first_computed_at":"2026-07-07T00:16:20.843646Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"EvoOtter: Evolutionary Reproduction Test Generator","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.SE","authors_text":"Avraham Shinnar, Jatin Ganhotra, Martin Hirzel, Toufique Ahmed","submitted_at":"2026-07-03T01:30:37Z","abstract_excerpt":"Before fixing an issue, it is useful to first reproduce it by generating a bug reproduction test (BRT). However, generating a BRT is itself a challenging task, because issue descriptions tend to be informal, making it difficult to determine whether a candidate BRT indeed fails for the reason in the issue. Prior work has attempted to tackle this problem via inference scaling, using large language models to generate many BRTs and patches, then using execution feedback to select and improve them. Unfortunately, this is expensive and the feedback is unreliable. This paper explores evolutionary pro"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2607.02854","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/2607.02854/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":"2607.02854","created_at":"2026-07-07T00:16:20.843733+00:00"},{"alias_kind":"arxiv_version","alias_value":"2607.02854v1","created_at":"2026-07-07T00:16:20.843733+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.02854","created_at":"2026-07-07T00:16:20.843733+00:00"},{"alias_kind":"pith_short_12","alias_value":"SGXTMIW2OCWC","created_at":"2026-07-07T00:16:20.843733+00:00"},{"alias_kind":"pith_short_16","alias_value":"SGXTMIW2OCWCSUE3","created_at":"2026-07-07T00:16:20.843733+00:00"},{"alias_kind":"pith_short_8","alias_value":"SGXTMIW2","created_at":"2026-07-07T00:16:20.843733+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/SGXTMIW2OCWCSUE3OAVK7ZLRDR","json":"https://pith.science/pith/SGXTMIW2OCWCSUE3OAVK7ZLRDR.json","graph_json":"https://pith.science/api/pith-number/SGXTMIW2OCWCSUE3OAVK7ZLRDR/graph.json","events_json":"https://pith.science/api/pith-number/SGXTMIW2OCWCSUE3OAVK7ZLRDR/events.json","paper":"https://pith.science/paper/SGXTMIW2"},"agent_actions":{"view_html":"https://pith.science/pith/SGXTMIW2OCWCSUE3OAVK7ZLRDR","download_json":"https://pith.science/pith/SGXTMIW2OCWCSUE3OAVK7ZLRDR.json","view_paper":"https://pith.science/paper/SGXTMIW2","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2607.02854&json=true","fetch_graph":"https://pith.science/api/pith-number/SGXTMIW2OCWCSUE3OAVK7ZLRDR/graph.json","fetch_events":"https://pith.science/api/pith-number/SGXTMIW2OCWCSUE3OAVK7ZLRDR/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/SGXTMIW2OCWCSUE3OAVK7ZLRDR/action/timestamp_anchor","attest_storage":"https://pith.science/pith/SGXTMIW2OCWCSUE3OAVK7ZLRDR/action/storage_attestation","attest_author":"https://pith.science/pith/SGXTMIW2OCWCSUE3OAVK7ZLRDR/action/author_attestation","sign_citation":"https://pith.science/pith/SGXTMIW2OCWCSUE3OAVK7ZLRDR/action/citation_signature","submit_replication":"https://pith.science/pith/SGXTMIW2OCWCSUE3OAVK7ZLRDR/action/replication_record"}},"created_at":"2026-07-07T00:16:20.843733+00:00","updated_at":"2026-07-07T00:16:20.843733+00:00"}