{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:XIZ37DTT3E5YHYZF22DHZDQLJA","short_pith_number":"pith:XIZ37DTT","canonical_record":{"source":{"id":"2605.16725","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.AI","submitted_at":"2026-05-16T00:18:22Z","cross_cats_sorted":[],"title_canon_sha256":"384c1aa69880624b23ac84fb58fd13d016a63bf9595c8bb40287786018492ac3","abstract_canon_sha256":"3bec55d75d813f3e7e727aab2bb3529e600a70edf771acacf2fe5d5ef4021952"},"schema_version":"1.0"},"canonical_sha256":"ba33bf8e73d93b83e325d6867c8e0b4817ff7a952183762f9e7bc9064a708d70","source":{"kind":"arxiv","id":"2605.16725","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.16725","created_at":"2026-05-20T00:02:38Z"},{"alias_kind":"arxiv_version","alias_value":"2605.16725v1","created_at":"2026-05-20T00:02:38Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.16725","created_at":"2026-05-20T00:02:38Z"},{"alias_kind":"pith_short_12","alias_value":"XIZ37DTT3E5Y","created_at":"2026-05-20T00:02:38Z"},{"alias_kind":"pith_short_16","alias_value":"XIZ37DTT3E5YHYZF","created_at":"2026-05-20T00:02:38Z"},{"alias_kind":"pith_short_8","alias_value":"XIZ37DTT","created_at":"2026-05-20T00:02:38Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:XIZ37DTT3E5YHYZF22DHZDQLJA","target":"record","payload":{"canonical_record":{"source":{"id":"2605.16725","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.AI","submitted_at":"2026-05-16T00:18:22Z","cross_cats_sorted":[],"title_canon_sha256":"384c1aa69880624b23ac84fb58fd13d016a63bf9595c8bb40287786018492ac3","abstract_canon_sha256":"3bec55d75d813f3e7e727aab2bb3529e600a70edf771acacf2fe5d5ef4021952"},"schema_version":"1.0"},"canonical_sha256":"ba33bf8e73d93b83e325d6867c8e0b4817ff7a952183762f9e7bc9064a708d70","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-20T00:02:38.601112Z","signature_b64":"PWTJ2Q1S1fgVg9hISXDjqF9updlunfW7VynFx/epGkSnXbxg+9wdXAlHwGJL1ZSKe9XQWwHChhgFv9UjefRbDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"ba33bf8e73d93b83e325d6867c8e0b4817ff7a952183762f9e7bc9064a708d70","last_reissued_at":"2026-05-20T00:02:38.600366Z","signature_status":"signed_v1","first_computed_at":"2026-05-20T00:02:38.600366Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2605.16725","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":"2026-05-20T00:02:38Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"/vmtkUElxiRw0dAp/7BeeC963/g3ZwPZoqHIO9f0zVGg3JzQLoLbR0LpbiB3lsSHdU+62LZA7bF4ZE+90C+aAQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-22T07:12:42.489313Z"},"content_sha256":"7f06dc79853d34909496c998c3f1df5dc6968653c69151571e5d1a7030fda8d9","schema_version":"1.0","event_id":"sha256:7f06dc79853d34909496c998c3f1df5dc6968653c69151571e5d1a7030fda8d9"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:XIZ37DTT3E5YHYZF22DHZDQLJA","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Baba in Wonderland: Online Self-Supervised Dynamics Discovery for Executable World Models","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"Alice learns executable world models by refining failed candidate updates into hypothesis classes","cross_cats":[],"primary_cat":"cs.AI","authors_text":"DongHeun Han, HyeongYeop Kang, SeongRae Noh, SeungWon Seo","submitted_at":"2026-05-16T00:18:22Z","abstract_excerpt":"Executable world models can be read, edited, executed, and reused for planning, but only if the program captures the environment's transition law rather than semantic shortcuts in its surface vocabulary. We study online executable world-model learning under prior misalignment, where an agent must induce state-dependent dynamics from interaction evidence alone, without rule descriptions, reward signals, or trustworthy lexical priors. We introduce Alice, a closed-loop system that treats failed candidate updates as structural signal: when a candidate explains a new transition but loses previously"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"Experiments show that Alice substantially improves executable world-model learning under prior misalignment, and ablations show that both class refinement and class-aware exploration contribute.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That failed candidate updates provide structural signal revealing dynamics the current program had conflated, and that refining these into hypothesis classes yields compact preservation counterexamples sufficient for effective updates.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"Alice uses preservation conflicts from failed candidate updates to create class-stratified hypotheses and guide exploration, improving executable world-model learning under prior misalignment.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Alice learns executable world models by refining failed candidate updates into hypothesis classes","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"18d32c9a64257b508721960e3333e98c78188215464d7f1f501131501466c0ab"},"source":{"id":"2605.16725","kind":"arxiv","version":1},"verdict":{"id":"b120f808-9e55-48fb-b74f-31384bc57a18","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-19T21:43:43.867594Z","strongest_claim":"Experiments show that Alice substantially improves executable world-model learning under prior misalignment, and ablations show that both class refinement and class-aware exploration contribute.","one_line_summary":"Alice uses preservation conflicts from failed candidate updates to create class-stratified hypotheses and guide exploration, improving executable world-model learning under prior misalignment.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That failed candidate updates provide structural signal revealing dynamics the current program had conflated, and that refining these into hypothesis classes yields compact preservation counterexamples sufficient for effective updates.","pith_extraction_headline":"Alice learns executable world models by refining failed candidate updates into hypothesis classes"},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2605.16725/integrity.json","findings":[],"available":true,"detectors_run":[{"name":"doi_title_agreement","ran_at":"2026-05-19T22:01:23.095741Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_compliance","ran_at":"2026-05-19T21:51:05.369648Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"claim_evidence","ran_at":"2026-05-19T19:01:56.347822Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"ai_meta_artifact","ran_at":"2026-05-19T18:33:26.473059Z","status":"skipped","version":"1.0.0","findings_count":0}],"snapshot_sha256":"156e17e80c91e0caefdb9c7b6c53721da4eff4bf08637909c036949660431ac0"},"references":{"count":37,"sample":[{"doi":"","year":2002,"title":"Never give up: Learning directed exploration strategies.arXiv preprint arXiv:2002.06038, 2020","work_id":"ca54caa0-72d9-4b25-b46f-b748b7ea4d4d","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2018,"title":"Exploration by random network distillation","work_id":"5a87fef6-96e2-4d5b-91ec-1a7c9a43cab9","ref_index":2,"cited_arxiv_id":"1810.12894","is_internal_anchor":true},{"doi":"","year":2024,"title":"Generating code world models with large language models guided by monte carlo tree search.Advances in Neural Information Processing Systems, 37:60429–60474, 2024","work_id":"bab6f6ad-a115-408f-8e1a-1691c5b77422","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":1901,"title":"Stanley, and Jeff Clune","work_id":"811ed60c-dea3-46b6-b28f-bf76eb844c38","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2018,"title":"Diversity is All You Need: Learning Skills without a Reward Function","work_id":"2ee80929-ae4d-4342-a8f3-ad970e340f75","ref_index":5,"cited_arxiv_id":"1802.06070","is_internal_anchor":true}],"resolved_work":37,"snapshot_sha256":"7eb0ce61604064a2092e161531553f0ca872818655d9d9d6c2ea497fb166227a","internal_anchors":6},"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":"b120f808-9e55-48fb-b74f-31384bc57a18"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-20T00:02:38Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"n1c/SWztkASrtnqezPwEBeIFXIiAkCfOW6/C8/oqejbSE8cC8zJVoRYrepjyErilTAHlt0lXhDQQj7OvKuZTAw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-22T07:12:42.490592Z"},"content_sha256":"28010c04b6d55f59b99912947158f7aeb6e2bad4ec12f26780f776a15e0bf5a4","schema_version":"1.0","event_id":"sha256:28010c04b6d55f59b99912947158f7aeb6e2bad4ec12f26780f776a15e0bf5a4"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/XIZ37DTT3E5YHYZF22DHZDQLJA/bundle.json","state_url":"https://pith.science/pith/XIZ37DTT3E5YHYZF22DHZDQLJA/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/XIZ37DTT3E5YHYZF22DHZDQLJA/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":"2026-05-22T07:12:42Z","links":{"resolver":"https://pith.science/pith/XIZ37DTT3E5YHYZF22DHZDQLJA","bundle":"https://pith.science/pith/XIZ37DTT3E5YHYZF22DHZDQLJA/bundle.json","state":"https://pith.science/pith/XIZ37DTT3E5YHYZF22DHZDQLJA/state.json","well_known_bundle":"https://pith.science/.well-known/pith/XIZ37DTT3E5YHYZF22DHZDQLJA/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:XIZ37DTT3E5YHYZF22DHZDQLJA","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":"3bec55d75d813f3e7e727aab2bb3529e600a70edf771acacf2fe5d5ef4021952","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.AI","submitted_at":"2026-05-16T00:18:22Z","title_canon_sha256":"384c1aa69880624b23ac84fb58fd13d016a63bf9595c8bb40287786018492ac3"},"schema_version":"1.0","source":{"id":"2605.16725","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.16725","created_at":"2026-05-20T00:02:38Z"},{"alias_kind":"arxiv_version","alias_value":"2605.16725v1","created_at":"2026-05-20T00:02:38Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.16725","created_at":"2026-05-20T00:02:38Z"},{"alias_kind":"pith_short_12","alias_value":"XIZ37DTT3E5Y","created_at":"2026-05-20T00:02:38Z"},{"alias_kind":"pith_short_16","alias_value":"XIZ37DTT3E5YHYZF","created_at":"2026-05-20T00:02:38Z"},{"alias_kind":"pith_short_8","alias_value":"XIZ37DTT","created_at":"2026-05-20T00:02:38Z"}],"graph_snapshots":[{"event_id":"sha256:28010c04b6d55f59b99912947158f7aeb6e2bad4ec12f26780f776a15e0bf5a4","target":"graph","created_at":"2026-05-20T00:02:38Z","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":4,"items":[{"attestation":"unclaimed","claim_id":"C1","kind":"strongest_claim","source":"verdict.strongest_claim","status":"machine_extracted","text":"Experiments show that Alice substantially improves executable world-model learning under prior misalignment, and ablations show that both class refinement and class-aware exploration contribute."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"That failed candidate updates provide structural signal revealing dynamics the current program had conflated, and that refining these into hypothesis classes yields compact preservation counterexamples sufficient for effective updates."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"Alice uses preservation conflicts from failed candidate updates to create class-stratified hypotheses and guide exploration, improving executable world-model learning under prior misalignment."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"Alice learns executable world models by refining failed candidate updates into hypothesis classes"}],"snapshot_sha256":"18d32c9a64257b508721960e3333e98c78188215464d7f1f501131501466c0ab"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[{"findings_count":0,"name":"doi_title_agreement","ran_at":"2026-05-19T22:01:23.095741Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"doi_compliance","ran_at":"2026-05-19T21:51:05.369648Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"claim_evidence","ran_at":"2026-05-19T19:01:56.347822Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"ai_meta_artifact","ran_at":"2026-05-19T18:33:26.473059Z","status":"skipped","version":"1.0.0"}],"endpoint":"/pith/2605.16725/integrity.json","findings":[],"snapshot_sha256":"156e17e80c91e0caefdb9c7b6c53721da4eff4bf08637909c036949660431ac0","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Executable world models can be read, edited, executed, and reused for planning, but only if the program captures the environment's transition law rather than semantic shortcuts in its surface vocabulary. We study online executable world-model learning under prior misalignment, where an agent must induce state-dependent dynamics from interaction evidence alone, without rule descriptions, reward signals, or trustworthy lexical priors. We introduce Alice, a closed-loop system that treats failed candidate updates as structural signal: when a candidate explains a new transition but loses previously","authors_text":"DongHeun Han, HyeongYeop Kang, SeongRae Noh, SeungWon Seo","cross_cats":[],"headline":"Alice learns executable world models by refining failed candidate updates into hypothesis classes","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.AI","submitted_at":"2026-05-16T00:18:22Z","title":"Baba in Wonderland: Online Self-Supervised Dynamics Discovery for Executable World Models"},"references":{"count":37,"internal_anchors":6,"resolved_work":37,"sample":[{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":1,"title":"Never give up: Learning directed exploration strategies.arXiv preprint arXiv:2002.06038, 2020","work_id":"ca54caa0-72d9-4b25-b46f-b748b7ea4d4d","year":2002},{"cited_arxiv_id":"1810.12894","doi":"","is_internal_anchor":true,"ref_index":2,"title":"Exploration by random network distillation","work_id":"5a87fef6-96e2-4d5b-91ec-1a7c9a43cab9","year":2018},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":3,"title":"Generating code world models with large language models guided by monte carlo tree search.Advances in Neural Information Processing Systems, 37:60429–60474, 2024","work_id":"bab6f6ad-a115-408f-8e1a-1691c5b77422","year":2024},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":4,"title":"Stanley, and Jeff Clune","work_id":"811ed60c-dea3-46b6-b28f-bf76eb844c38","year":1901},{"cited_arxiv_id":"1802.06070","doi":"","is_internal_anchor":true,"ref_index":5,"title":"Diversity is All You Need: Learning Skills without a Reward Function","work_id":"2ee80929-ae4d-4342-a8f3-ad970e340f75","year":2018}],"snapshot_sha256":"7eb0ce61604064a2092e161531553f0ca872818655d9d9d6c2ea497fb166227a"},"source":{"id":"2605.16725","kind":"arxiv","version":1},"verdict":{"created_at":"2026-05-19T21:43:43.867594Z","id":"b120f808-9e55-48fb-b74f-31384bc57a18","model_set":{"reader":"grok-4.3"},"one_line_summary":"Alice uses preservation conflicts from failed candidate updates to create class-stratified hypotheses and guide exploration, improving executable world-model learning under prior misalignment.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"Alice learns executable world models by refining failed candidate updates into hypothesis classes","strongest_claim":"Experiments show that Alice substantially improves executable world-model learning under prior misalignment, and ablations show that both class refinement and class-aware exploration contribute.","weakest_assumption":"That failed candidate updates provide structural signal revealing dynamics the current program had conflated, and that refining these into hypothesis classes yields compact preservation counterexamples sufficient for effective updates."}},"verdict_id":"b120f808-9e55-48fb-b74f-31384bc57a18"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:7f06dc79853d34909496c998c3f1df5dc6968653c69151571e5d1a7030fda8d9","target":"record","created_at":"2026-05-20T00:02:38Z","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":"3bec55d75d813f3e7e727aab2bb3529e600a70edf771acacf2fe5d5ef4021952","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.AI","submitted_at":"2026-05-16T00:18:22Z","title_canon_sha256":"384c1aa69880624b23ac84fb58fd13d016a63bf9595c8bb40287786018492ac3"},"schema_version":"1.0","source":{"id":"2605.16725","kind":"arxiv","version":1}},"canonical_sha256":"ba33bf8e73d93b83e325d6867c8e0b4817ff7a952183762f9e7bc9064a708d70","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"ba33bf8e73d93b83e325d6867c8e0b4817ff7a952183762f9e7bc9064a708d70","first_computed_at":"2026-05-20T00:02:38.600366Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-20T00:02:38.600366Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"PWTJ2Q1S1fgVg9hISXDjqF9updlunfW7VynFx/epGkSnXbxg+9wdXAlHwGJL1ZSKe9XQWwHChhgFv9UjefRbDQ==","signature_status":"signed_v1","signed_at":"2026-05-20T00:02:38.601112Z","signed_message":"canonical_sha256_bytes"},"source_id":"2605.16725","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:7f06dc79853d34909496c998c3f1df5dc6968653c69151571e5d1a7030fda8d9","sha256:28010c04b6d55f59b99912947158f7aeb6e2bad4ec12f26780f776a15e0bf5a4"],"state_sha256":"22b1c31ae35a0b7d59d44a05bb4519991af5133bde9e04a02d260eb6deca83f6"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"qy/BJm3gMY3yetWam6OJMh6BLYiGmQX8rYRJpr5ZmQZtRxOniGnI/y4PmxNe/azAQZF0L8r29jGt/XwNrSwSCQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-05-22T07:12:42.495394Z","bundle_sha256":"0f203dd886c996c6f9c0f38f9b42f92c51695d2264e3fd10a6296ba6fa335db3"}}