{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:DV5O44BXXN6H6B3N3CIE63GMOV","short_pith_number":"pith:DV5O44BX","canonical_record":{"source":{"id":"2507.03724","kind":"arxiv","version":4},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-07-04T17:21:46Z","cross_cats_sorted":[],"title_canon_sha256":"aef56d57a3080f496ff97f65475109cb180e40c9b80c4633af062489141ed88b","abstract_canon_sha256":"4ebf7b5154c4fd6be78e3094702f207fcc5551ef5798dad4fa6dba851865f272"},"schema_version":"1.0"},"canonical_sha256":"1d7aee7037bb7c7f076dd8904f6ccc757ef2a0d7f671d93e9f50ab76b12603bc","source":{"kind":"arxiv","id":"2507.03724","version":4},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2507.03724","created_at":"2026-05-17T23:38:53Z"},{"alias_kind":"arxiv_version","alias_value":"2507.03724v4","created_at":"2026-05-17T23:38:53Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2507.03724","created_at":"2026-05-17T23:38:53Z"},{"alias_kind":"pith_short_12","alias_value":"DV5O44BXXN6H","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_16","alias_value":"DV5O44BXXN6H6B3N","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_8","alias_value":"DV5O44BX","created_at":"2026-05-18T12:33:37Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:DV5O44BXXN6H6B3N3CIE63GMOV","target":"record","payload":{"canonical_record":{"source":{"id":"2507.03724","kind":"arxiv","version":4},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-07-04T17:21:46Z","cross_cats_sorted":[],"title_canon_sha256":"aef56d57a3080f496ff97f65475109cb180e40c9b80c4633af062489141ed88b","abstract_canon_sha256":"4ebf7b5154c4fd6be78e3094702f207fcc5551ef5798dad4fa6dba851865f272"},"schema_version":"1.0"},"canonical_sha256":"1d7aee7037bb7c7f076dd8904f6ccc757ef2a0d7f671d93e9f50ab76b12603bc","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-17T23:38:53.039878Z","signature_b64":"PcnkEgO8L5iY9eh/KejZb5UPOLgZFvOI/dbViCuv1BQ11B5+bGLqNHIphr+z/HDBQbtwenM8jE/UTpI6sH1lDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"1d7aee7037bb7c7f076dd8904f6ccc757ef2a0d7f671d93e9f50ab76b12603bc","last_reissued_at":"2026-05-17T23:38:53.039357Z","signature_status":"signed_v1","first_computed_at":"2026-05-17T23:38:53.039357Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2507.03724","source_version":4,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-17T23:38:53Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"QZ7NjV2AaC0uMc2Nc8e+JbDQMQxnY2uV3CGLKf7uUQayriP3JpfLT8WuGXrMLy5K2/tIp/YVHhPgwQkaqZrrAA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-21T05:42:18.934234Z"},"content_sha256":"236fd08061e3d6f3e7ac89846833a832d7e2e54e482b9a3e80961f1365f20091","schema_version":"1.0","event_id":"sha256:236fd08061e3d6f3e7ac89846833a832d7e2e54e482b9a3e80961f1365f20091"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:DV5O44BXXN6H6B3N3CIE63GMOV","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"MemOS: A Memory OS for AI System","license":"http://creativecommons.org/licenses/by/4.0/","headline":"MemOS proposes a memory operating system that unifies plaintext, activation-based, and parameter memories in LLMs through MemCubes.","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Bo Tang, Boyu Chen, Chen Tang, Chenyang Xi, Chunyu Li, Ding Chen, Feiyu Xiong, Hanyu Wang, Haofen Wang, Hao Wu, Hongkang Yang, Huajun Chen, Huayi Lai, Jiahao Huo, Jiawei Yang, Jihao Zhao, Junchi Yan, Kai Chen, Kehang Li, Linfeng Zhang, Mingchuan Yang, Ningyu Zhang, Peng Liu, Pengyuan Wang, Qingchen Yu, Shichao Song, Siheng Chen, Simin Niu, Tianyi Chen, Tong Xu, Wei Xu, Wentao Zhang, Yezhaohui Wang, Zehao Lin, Zhaoxin Fan, Zhengren Wang, Zhen Tao, Zhi-Qin John Xu, Zhiyu Li","submitted_at":"2025-07-04T17:21:46Z","abstract_excerpt":"Large Language Models (LLMs) have become an essential infrastructure for Artificial General Intelligence (AGI), yet their lack of well-defined memory management systems hinders the development of long-context reasoning, continual personalization, and knowledge consistency.Existing models mainly rely on static parameters and short-lived contextual states, limiting their ability to track user preferences or update knowledge over extended periods.While Retrieval-Augmented Generation (RAG) introduces external knowledge in plain text, it remains a stateless workaround without lifecycle control or i"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"MemOS establishes a memory-centric system framework that brings controllability, plasticity, and evolvability to LLMs, laying the foundation for continual learning and personalized modeling.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"The assumption that a unified memory layer using MemCubes can be practically implemented to bridge retrieval and parameter-based learning while delivering the claimed reductions in cost and gains in consistency, without detailed mechanisms or validation provided.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"MemOS introduces a unified memory management framework for LLMs using MemCubes to handle and evolve different memory types for improved controllability and evolvability.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"MemOS proposes a memory operating system that unifies plaintext, activation-based, and parameter memories in LLMs through MemCubes.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"9e9423d84adbf8d87b2f1d6a9553ab0fad82ea882149504d3917781989842711"},"source":{"id":"2507.03724","kind":"arxiv","version":4},"verdict":{"id":"83754e74-f259-4947-a556-c0d23f336df3","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-15T08:15:28.772869Z","strongest_claim":"MemOS establishes a memory-centric system framework that brings controllability, plasticity, and evolvability to LLMs, laying the foundation for continual learning and personalized modeling.","one_line_summary":"MemOS introduces a unified memory management framework for LLMs using MemCubes to handle and evolve different memory types for improved controllability and evolvability.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"The assumption that a unified memory layer using MemCubes can be practically implemented to bridge retrieval and parameter-based learning while delivering the claimed reductions in cost and gains in consistency, without detailed mechanisms or validation provided.","pith_extraction_headline":"MemOS proposes a memory operating system that unifies plaintext, activation-based, and parameter memories in LLMs through MemCubes."},"references":{"count":106,"sample":[{"doi":"","year":2024,"title":"Memory3: Language modeling with explicit memory.Journal of Machine Learning, 3(3):300–346, January 2024","work_id":"51870a39-7f08-4cec-b2d5-6f6b56f6fa9b","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2023,"title":"A Survey of Large Language Models","work_id":"de1b42b5-4a0a-4b1f-8c78-1f7fe21be6c9","ref_index":2,"cited_arxiv_id":"2303.18223","is_internal_anchor":true},{"doi":"","year":2023,"title":"Codet5+: Open code large language models for code understanding and generation","work_id":"9a1f53cf-da37-4377-9707-c043ef9e5710","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2024,"title":"From linguistic giants to sensory maestros: A survey on cross-modal reasoning with large language models.arXiv preprint arXiv:2409.18996, 2024","work_id":"5966cc14-debd-4c6f-bfae-c8c86ab6a993","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2024,"title":"Retrieval-Augmented Generation for AI-Generated Content: A Survey","work_id":"5440e856-5c59-44cb-8ea8-f1c3591c7425","ref_index":5,"cited_arxiv_id":"2402.19473","is_internal_anchor":true}],"resolved_work":106,"snapshot_sha256":"a89df39b9b5c5d4f6c48fc37e1a5db533bace9a04da67430e6fd3706b56ff512","internal_anchors":27},"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":"83754e74-f259-4947-a556-c0d23f336df3"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-17T23:38:53Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"0GFOgCv7jwCvB5deuLSRsRvh3NnHZJEtIagsbHeJy4S345WOrnunBzbS+HY5ny8w7XpuwCOQcNECySA7q0gADw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-21T05:42:18.936347Z"},"content_sha256":"cadc5a09fb4a690be80636900e9f0bac9f57a3ddb8767357f0fe2a34636032b5","schema_version":"1.0","event_id":"sha256:cadc5a09fb4a690be80636900e9f0bac9f57a3ddb8767357f0fe2a34636032b5"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/DV5O44BXXN6H6B3N3CIE63GMOV/bundle.json","state_url":"https://pith.science/pith/DV5O44BXXN6H6B3N3CIE63GMOV/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/DV5O44BXXN6H6B3N3CIE63GMOV/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":"2026-08-21T05:42:18Z","links":{"resolver":"https://pith.science/pith/DV5O44BXXN6H6B3N3CIE63GMOV","bundle":"https://pith.science/pith/DV5O44BXXN6H6B3N3CIE63GMOV/bundle.json","state":"https://pith.science/pith/DV5O44BXXN6H6B3N3CIE63GMOV/state.json","well_known_bundle":"https://pith.science/.well-known/pith/DV5O44BXXN6H6B3N3CIE63GMOV/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:DV5O44BXXN6H6B3N3CIE63GMOV","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":"4ebf7b5154c4fd6be78e3094702f207fcc5551ef5798dad4fa6dba851865f272","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-07-04T17:21:46Z","title_canon_sha256":"aef56d57a3080f496ff97f65475109cb180e40c9b80c4633af062489141ed88b"},"schema_version":"1.0","source":{"id":"2507.03724","kind":"arxiv","version":4}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2507.03724","created_at":"2026-05-17T23:38:53Z"},{"alias_kind":"arxiv_version","alias_value":"2507.03724v4","created_at":"2026-05-17T23:38:53Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2507.03724","created_at":"2026-05-17T23:38:53Z"},{"alias_kind":"pith_short_12","alias_value":"DV5O44BXXN6H","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_16","alias_value":"DV5O44BXXN6H6B3N","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_8","alias_value":"DV5O44BX","created_at":"2026-05-18T12:33:37Z"}],"graph_snapshots":[{"event_id":"sha256:cadc5a09fb4a690be80636900e9f0bac9f57a3ddb8767357f0fe2a34636032b5","target":"graph","created_at":"2026-05-17T23:38:53Z","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":"MemOS establishes a memory-centric system framework that brings controllability, plasticity, and evolvability to LLMs, laying the foundation for continual learning and personalized modeling."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"The assumption that a unified memory layer using MemCubes can be practically implemented to bridge retrieval and parameter-based learning while delivering the claimed reductions in cost and gains in consistency, without detailed mechanisms or validation provided."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"MemOS introduces a unified memory management framework for LLMs using MemCubes to handle and evolve different memory types for improved controllability and evolvability."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"MemOS proposes a memory operating system that unifies plaintext, activation-based, and parameter memories in LLMs through MemCubes."}],"snapshot_sha256":"9e9423d84adbf8d87b2f1d6a9553ab0fad82ea882149504d3917781989842711"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"paper":{"abstract_excerpt":"Large Language Models (LLMs) have become an essential infrastructure for Artificial General Intelligence (AGI), yet their lack of well-defined memory management systems hinders the development of long-context reasoning, continual personalization, and knowledge consistency.Existing models mainly rely on static parameters and short-lived contextual states, limiting their ability to track user preferences or update knowledge over extended periods.While Retrieval-Augmented Generation (RAG) introduces external knowledge in plain text, it remains a stateless workaround without lifecycle control or i","authors_text":"Bo Tang, Boyu Chen, Chen Tang, Chenyang Xi, Chunyu Li, Ding Chen, Feiyu Xiong, Hanyu Wang, Haofen Wang, Hao Wu, Hongkang Yang, Huajun Chen, Huayi Lai, Jiahao Huo, Jiawei Yang, Jihao Zhao, Junchi Yan, Kai Chen, Kehang Li, Linfeng Zhang, Mingchuan Yang, Ningyu Zhang, Peng Liu, Pengyuan Wang, Qingchen Yu, Shichao Song, Siheng Chen, Simin Niu, Tianyi Chen, Tong Xu, Wei Xu, Wentao Zhang, Yezhaohui Wang, Zehao Lin, Zhaoxin Fan, Zhengren Wang, Zhen Tao, Zhi-Qin John Xu, Zhiyu Li","cross_cats":[],"headline":"MemOS proposes a memory operating system that unifies plaintext, activation-based, and parameter memories in LLMs through MemCubes.","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-07-04T17:21:46Z","title":"MemOS: A Memory OS for AI System"},"references":{"count":106,"internal_anchors":27,"resolved_work":106,"sample":[{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":1,"title":"Memory3: Language modeling with explicit memory.Journal of Machine Learning, 3(3):300–346, January 2024","work_id":"51870a39-7f08-4cec-b2d5-6f6b56f6fa9b","year":2024},{"cited_arxiv_id":"2303.18223","doi":"","is_internal_anchor":true,"ref_index":2,"title":"A Survey of Large Language Models","work_id":"de1b42b5-4a0a-4b1f-8c78-1f7fe21be6c9","year":2023},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":3,"title":"Codet5+: Open code large language models for code understanding and generation","work_id":"9a1f53cf-da37-4377-9707-c043ef9e5710","year":2023},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":4,"title":"From linguistic giants to sensory maestros: A survey on cross-modal reasoning with large language models.arXiv preprint arXiv:2409.18996, 2024","work_id":"5966cc14-debd-4c6f-bfae-c8c86ab6a993","year":2024},{"cited_arxiv_id":"2402.19473","doi":"","is_internal_anchor":true,"ref_index":5,"title":"Retrieval-Augmented Generation for AI-Generated Content: A Survey","work_id":"5440e856-5c59-44cb-8ea8-f1c3591c7425","year":2024}],"snapshot_sha256":"a89df39b9b5c5d4f6c48fc37e1a5db533bace9a04da67430e6fd3706b56ff512"},"source":{"id":"2507.03724","kind":"arxiv","version":4},"verdict":{"created_at":"2026-05-15T08:15:28.772869Z","id":"83754e74-f259-4947-a556-c0d23f336df3","model_set":{"reader":"grok-4.3"},"one_line_summary":"MemOS introduces a unified memory management framework for LLMs using MemCubes to handle and evolve different memory types for improved controllability and evolvability.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"MemOS proposes a memory operating system that unifies plaintext, activation-based, and parameter memories in LLMs through MemCubes.","strongest_claim":"MemOS establishes a memory-centric system framework that brings controllability, plasticity, and evolvability to LLMs, laying the foundation for continual learning and personalized modeling.","weakest_assumption":"The assumption that a unified memory layer using MemCubes can be practically implemented to bridge retrieval and parameter-based learning while delivering the claimed reductions in cost and gains in consistency, without detailed mechanisms or validation provided."}},"verdict_id":"83754e74-f259-4947-a556-c0d23f336df3"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:236fd08061e3d6f3e7ac89846833a832d7e2e54e482b9a3e80961f1365f20091","target":"record","created_at":"2026-05-17T23:38:53Z","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":"4ebf7b5154c4fd6be78e3094702f207fcc5551ef5798dad4fa6dba851865f272","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-07-04T17:21:46Z","title_canon_sha256":"aef56d57a3080f496ff97f65475109cb180e40c9b80c4633af062489141ed88b"},"schema_version":"1.0","source":{"id":"2507.03724","kind":"arxiv","version":4}},"canonical_sha256":"1d7aee7037bb7c7f076dd8904f6ccc757ef2a0d7f671d93e9f50ab76b12603bc","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"1d7aee7037bb7c7f076dd8904f6ccc757ef2a0d7f671d93e9f50ab76b12603bc","first_computed_at":"2026-05-17T23:38:53.039357Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-17T23:38:53.039357Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"PcnkEgO8L5iY9eh/KejZb5UPOLgZFvOI/dbViCuv1BQ11B5+bGLqNHIphr+z/HDBQbtwenM8jE/UTpI6sH1lDQ==","signature_status":"signed_v1","signed_at":"2026-05-17T23:38:53.039878Z","signed_message":"canonical_sha256_bytes"},"source_id":"2507.03724","source_kind":"arxiv","source_version":4}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:236fd08061e3d6f3e7ac89846833a832d7e2e54e482b9a3e80961f1365f20091","sha256:cadc5a09fb4a690be80636900e9f0bac9f57a3ddb8767357f0fe2a34636032b5"],"state_sha256":"eda7d68cbf4f7326603248f48751082e0ce6fabf718197d5e7503946a6a26bd5"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"RJmaZiNL+6G5RkpFQDmPXtR6XEmAGWastcEVzpZc1M7RPBDkr50XWoZu6hHHMGQJs7IKgdnB4QXIZsgQe/A7AA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-21T05:42:18.989570Z","bundle_sha256":"24164e0ae475a5d595187481ef4243f886c9f811cf59c6bc48dfc50498aef844"}}