{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:RFZTG4EWXJSEHH4VCKNEVWOLRE","short_pith_number":"pith:RFZTG4EW","canonical_record":{"source":{"id":"2502.18356","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-02-25T16:45:08Z","cross_cats_sorted":[],"title_canon_sha256":"f721c34984276127962f9b6d8b113ec033a8a4ba8ba9ab55a17ed7dad4dc487b","abstract_canon_sha256":"191f20e35d5d7b039098a50225e481c9bae488ccc91d290a7e3686da1d2d38ef"},"schema_version":"1.0"},"canonical_sha256":"8973337096ba64439f95129a4ad9cb893422fa74d48f1b70928aece7f0dbd731","source":{"kind":"arxiv","id":"2502.18356","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2502.18356","created_at":"2026-07-05T10:19:52Z"},{"alias_kind":"arxiv_version","alias_value":"2502.18356v1","created_at":"2026-07-05T10:19:52Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2502.18356","created_at":"2026-07-05T10:19:52Z"},{"alias_kind":"pith_short_12","alias_value":"RFZTG4EWXJSE","created_at":"2026-07-05T10:19:52Z"},{"alias_kind":"pith_short_16","alias_value":"RFZTG4EWXJSEHH4V","created_at":"2026-07-05T10:19:52Z"},{"alias_kind":"pith_short_8","alias_value":"RFZTG4EW","created_at":"2026-07-05T10:19:52Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:RFZTG4EWXJSEHH4VCKNEVWOLRE","target":"record","payload":{"canonical_record":{"source":{"id":"2502.18356","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-02-25T16:45:08Z","cross_cats_sorted":[],"title_canon_sha256":"f721c34984276127962f9b6d8b113ec033a8a4ba8ba9ab55a17ed7dad4dc487b","abstract_canon_sha256":"191f20e35d5d7b039098a50225e481c9bae488ccc91d290a7e3686da1d2d38ef"},"schema_version":"1.0"},"canonical_sha256":"8973337096ba64439f95129a4ad9cb893422fa74d48f1b70928aece7f0dbd731","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:19:52.660253Z","signature_b64":"EjJHwrND759g/JTJTntobF/M3j4VuPan69RkX2SvcW/hfMLuZTJAHYAI7mtbjS2A9+LOvG1YbrpXGaOv73BtDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"8973337096ba64439f95129a4ad9cb893422fa74d48f1b70928aece7f0dbd731","last_reissued_at":"2026-07-05T10:19:52.659775Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:19:52.659775Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2502.18356","source_version":1,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T10:19:52Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"D8c1rxQWDjmksezXToR/BaiX37A9MZTvGWzBN/v5ULSqODy4N7giGJAeCEOXRLL6EJT+er9pcsxT2QKosjhHCw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-08T16:46:10.674653Z"},"content_sha256":"0dee2b7e8ed01682aa4e5c1be0e249382db5e58f1b00b9134699d38bb20d6c8e","schema_version":"1.0","event_id":"sha256:0dee2b7e8ed01682aa4e5c1be0e249382db5e58f1b00b9134699d38bb20d6c8e"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:RFZTG4EWXJSEHH4VCKNEVWOLRE","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"WebGames: Challenging General-Purpose Web-Browsing AI Agents","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Alex J. Chan, Andy Toulis, Filippos Christianos, Fraser Greenlee, George Thomas, Jikun Kang, Marvin Purtorab, Wenqi Wu","submitted_at":"2025-02-25T16:45:08Z","abstract_excerpt":"We introduce WebGames, a comprehensive benchmark suite designed to evaluate general-purpose web-browsing AI agents through a collection of 50+ interactive challenges. These challenges are specifically crafted to be straightforward for humans while systematically testing the limitations of current AI systems across fundamental browser interactions, advanced input processing, cognitive tasks, workflow automation, and interactive entertainment. Our framework eliminates external dependencies through a hermetic testing environment, ensuring reproducible evaluation with verifiable ground-truth solut"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2502.18356","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/2502.18356/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"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T10:19:52Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"T6T7bcbhtm2UZQP8ZBppMlsVM9pcmdgtga4XC1tEaqBdaVxb3V80bzohlm/tEQXKWJmShY3fZzE9DSA8ifimDQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-08T16:46:10.675196Z"},"content_sha256":"a5c87a8bdce69da5d4a563b5009067e04876810748d26459e65b11140115f218","schema_version":"1.0","event_id":"sha256:a5c87a8bdce69da5d4a563b5009067e04876810748d26459e65b11140115f218"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/RFZTG4EWXJSEHH4VCKNEVWOLRE/bundle.json","state_url":"https://pith.science/pith/RFZTG4EWXJSEHH4VCKNEVWOLRE/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/RFZTG4EWXJSEHH4VCKNEVWOLRE/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-08-08T16:46:10Z","links":{"resolver":"https://pith.science/pith/RFZTG4EWXJSEHH4VCKNEVWOLRE","bundle":"https://pith.science/pith/RFZTG4EWXJSEHH4VCKNEVWOLRE/bundle.json","state":"https://pith.science/pith/RFZTG4EWXJSEHH4VCKNEVWOLRE/state.json","well_known_bundle":"https://pith.science/.well-known/pith/RFZTG4EWXJSEHH4VCKNEVWOLRE/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:RFZTG4EWXJSEHH4VCKNEVWOLRE","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"191f20e35d5d7b039098a50225e481c9bae488ccc91d290a7e3686da1d2d38ef","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-02-25T16:45:08Z","title_canon_sha256":"f721c34984276127962f9b6d8b113ec033a8a4ba8ba9ab55a17ed7dad4dc487b"},"schema_version":"1.0","source":{"id":"2502.18356","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2502.18356","created_at":"2026-07-05T10:19:52Z"},{"alias_kind":"arxiv_version","alias_value":"2502.18356v1","created_at":"2026-07-05T10:19:52Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2502.18356","created_at":"2026-07-05T10:19:52Z"},{"alias_kind":"pith_short_12","alias_value":"RFZTG4EWXJSE","created_at":"2026-07-05T10:19:52Z"},{"alias_kind":"pith_short_16","alias_value":"RFZTG4EWXJSEHH4V","created_at":"2026-07-05T10:19:52Z"},{"alias_kind":"pith_short_8","alias_value":"RFZTG4EW","created_at":"2026-07-05T10:19:52Z"}],"graph_snapshots":[{"event_id":"sha256:a5c87a8bdce69da5d4a563b5009067e04876810748d26459e65b11140115f218","target":"graph","created_at":"2026-07-05T10:19:52Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2502.18356/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"We introduce WebGames, a comprehensive benchmark suite designed to evaluate general-purpose web-browsing AI agents through a collection of 50+ interactive challenges. These challenges are specifically crafted to be straightforward for humans while systematically testing the limitations of current AI systems across fundamental browser interactions, advanced input processing, cognitive tasks, workflow automation, and interactive entertainment. Our framework eliminates external dependencies through a hermetic testing environment, ensuring reproducible evaluation with verifiable ground-truth solut","authors_text":"Alex J. Chan, Andy Toulis, Filippos Christianos, Fraser Greenlee, George Thomas, Jikun Kang, Marvin Purtorab, Wenqi Wu","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-02-25T16:45:08Z","title":"WebGames: Challenging General-Purpose Web-Browsing AI Agents"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2502.18356","kind":"arxiv","version":1},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:0dee2b7e8ed01682aa4e5c1be0e249382db5e58f1b00b9134699d38bb20d6c8e","target":"record","created_at":"2026-07-05T10:19:52Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"191f20e35d5d7b039098a50225e481c9bae488ccc91d290a7e3686da1d2d38ef","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-02-25T16:45:08Z","title_canon_sha256":"f721c34984276127962f9b6d8b113ec033a8a4ba8ba9ab55a17ed7dad4dc487b"},"schema_version":"1.0","source":{"id":"2502.18356","kind":"arxiv","version":1}},"canonical_sha256":"8973337096ba64439f95129a4ad9cb893422fa74d48f1b70928aece7f0dbd731","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"8973337096ba64439f95129a4ad9cb893422fa74d48f1b70928aece7f0dbd731","first_computed_at":"2026-07-05T10:19:52.659775Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T10:19:52.659775Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"EjJHwrND759g/JTJTntobF/M3j4VuPan69RkX2SvcW/hfMLuZTJAHYAI7mtbjS2A9+LOvG1YbrpXGaOv73BtDw==","signature_status":"signed_v1","signed_at":"2026-07-05T10:19:52.660253Z","signed_message":"canonical_sha256_bytes"},"source_id":"2502.18356","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:0dee2b7e8ed01682aa4e5c1be0e249382db5e58f1b00b9134699d38bb20d6c8e","sha256:a5c87a8bdce69da5d4a563b5009067e04876810748d26459e65b11140115f218"],"state_sha256":"3381f2a3cf8f85ab0d66f630f5efe774bd8760f85038d505a50addb65afc5d6a"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"KYjMDiFJugoBku9agGwBR+c4mKz8xDewplYSnIMte2HnxC+aax1x9mYsplQ5maUqN3LRobsl3rFnMpAmXUO8Cg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-08T16:46:10.679312Z","bundle_sha256":"7de04cae924be011269955894ac5125c3e11e00fc8d4c72dc9201494a1e369ad"}}