{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:EE2J7NHHPUMEKMM4DCV6Q4LWHK","short_pith_number":"pith:EE2J7NHH","schema_version":"1.0","canonical_sha256":"21349fb4e77d1845319c18abe871763aa9456b772f783e66da7cf13b3ab52931","source":{"kind":"arxiv","id":"2605.17637","version":1},"attestation_state":"computed","paper":{"title":"WebGameBench: Requirement-to-Application Evaluation for Coding Agents via Browser-Native Games","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Daxiang Dong, Guoliang You, Haoran Wang, Haotian Zhao, Jianmin Wu, Jingnan Gu, Mingyang Dai, Tianlun, Tianshu Zhu, Wenyu Zhang, Xiaoxuan Tang","submitted_at":"2026-05-17T20:07:12Z","abstract_excerpt":"Coding agents are increasingly used as application builders, yet many evaluations still focus on source code, repository-level tests, or intermediate traces rather than the delivered application. We introduce WebGameBench, a requirement-to-application benchmark that evaluates whether coding agents can turn a frozen Structured WebGame Specification into a browser-accessible game. Browser-native games provide a compact but behavior-dense testbed: even simple games require coordinated input handling, spatial mapping, rule execution, state transitions, terminal conditions, restart behavior, and vi"},"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":"2605.17637","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2026-05-17T20:07:12Z","cross_cats_sorted":[],"title_canon_sha256":"48167c873137a1d2b9bc41c8c89f70ec595c17cf36771ac007ee1f498356d287","abstract_canon_sha256":"9999d7e83c1ba616c29727dbc12d6aa8090a3f64a596ee6b0d068024b35f68bf"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-20T00:04:49.887991Z","signature_b64":"A0ZVWJEPkky/4tcbyZxqvyfvstmaBFxUktmuDPW3BWVtrza9GyJ4U/LkjOqZzLvD3Wikg2N173bGT2KPqSrwBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"21349fb4e77d1845319c18abe871763aa9456b772f783e66da7cf13b3ab52931","last_reissued_at":"2026-05-20T00:04:49.886954Z","signature_status":"signed_v1","first_computed_at":"2026-05-20T00:04:49.886954Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"WebGameBench: Requirement-to-Application Evaluation for Coding Agents via Browser-Native Games","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Daxiang Dong, Guoliang You, Haoran Wang, Haotian Zhao, Jianmin Wu, Jingnan Gu, Mingyang Dai, Tianlun, Tianshu Zhu, Wenyu Zhang, Xiaoxuan Tang","submitted_at":"2026-05-17T20:07:12Z","abstract_excerpt":"Coding agents are increasingly used as application builders, yet many evaluations still focus on source code, repository-level tests, or intermediate traces rather than the delivered application. We introduce WebGameBench, a requirement-to-application benchmark that evaluates whether coding agents can turn a frozen Structured WebGame Specification into a browser-accessible game. Browser-native games provide a compact but behavior-dense testbed: even simple games require coordinated input handling, spatial mapping, rule execution, state transitions, terminal conditions, restart behavior, and vi"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.17637","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/2605.17637/integrity.json","findings":[],"available":true,"detectors_run":[{"name":"cited_work_retraction","ran_at":"2026-05-19T22:51:57.287241Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"ai_meta_artifact","ran_at":"2026-05-19T21:33:23.555049Z","status":"skipped","version":"1.0.0","findings_count":0},{"name":"claim_evidence","ran_at":"2026-05-19T21:21:57.477577Z","status":"completed","version":"1.0.0","findings_count":0}],"snapshot_sha256":"11e501630281cfe3799708531605edfc8656fb9d8a9dbf05e99f5eee05e85e20"},"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":"2605.17637","created_at":"2026-05-20T00:04:49.887105+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.17637v1","created_at":"2026-05-20T00:04:49.887105+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.17637","created_at":"2026-05-20T00:04:49.887105+00:00"},{"alias_kind":"pith_short_12","alias_value":"EE2J7NHHPUME","created_at":"2026-05-20T00:04:49.887105+00:00"},{"alias_kind":"pith_short_16","alias_value":"EE2J7NHHPUMEKMM4","created_at":"2026-05-20T00:04:49.887105+00:00"},{"alias_kind":"pith_short_8","alias_value":"EE2J7NHH","created_at":"2026-05-20T00:04:49.887105+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/EE2J7NHHPUMEKMM4DCV6Q4LWHK","json":"https://pith.science/pith/EE2J7NHHPUMEKMM4DCV6Q4LWHK.json","graph_json":"https://pith.science/api/pith-number/EE2J7NHHPUMEKMM4DCV6Q4LWHK/graph.json","events_json":"https://pith.science/api/pith-number/EE2J7NHHPUMEKMM4DCV6Q4LWHK/events.json","paper":"https://pith.science/paper/EE2J7NHH"},"agent_actions":{"view_html":"https://pith.science/pith/EE2J7NHHPUMEKMM4DCV6Q4LWHK","download_json":"https://pith.science/pith/EE2J7NHHPUMEKMM4DCV6Q4LWHK.json","view_paper":"https://pith.science/paper/EE2J7NHH","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.17637&json=true","fetch_graph":"https://pith.science/api/pith-number/EE2J7NHHPUMEKMM4DCV6Q4LWHK/graph.json","fetch_events":"https://pith.science/api/pith-number/EE2J7NHHPUMEKMM4DCV6Q4LWHK/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/EE2J7NHHPUMEKMM4DCV6Q4LWHK/action/timestamp_anchor","attest_storage":"https://pith.science/pith/EE2J7NHHPUMEKMM4DCV6Q4LWHK/action/storage_attestation","attest_author":"https://pith.science/pith/EE2J7NHHPUMEKMM4DCV6Q4LWHK/action/author_attestation","sign_citation":"https://pith.science/pith/EE2J7NHHPUMEKMM4DCV6Q4LWHK/action/citation_signature","submit_replication":"https://pith.science/pith/EE2J7NHHPUMEKMM4DCV6Q4LWHK/action/replication_record"}},"created_at":"2026-05-20T00:04:49.887105+00:00","updated_at":"2026-05-20T00:04:49.887105+00:00"}