{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:5OVVRCK3QAVKBZI6OERYJRBOVH","short_pith_number":"pith:5OVVRCK3","canonical_record":{"source":{"id":"2605.16857","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2026-05-16T07:41:31Z","cross_cats_sorted":[],"title_canon_sha256":"3faf8d33f20e447ff83f46d9344b3ffead9f7bbd8baf6157d7d9f123318951b4","abstract_canon_sha256":"382ab370292499827e39db0e9bec48cbe2f9635e8d4c452a1c677b52a8f58039"},"schema_version":"1.0"},"canonical_sha256":"ebab58895b802aa0e51e712384c42ea9e1cea3f8ed467c73aad54eb7aa66cabf","source":{"kind":"arxiv","id":"2605.16857","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.16857","created_at":"2026-05-20T00:03:26Z"},{"alias_kind":"arxiv_version","alias_value":"2605.16857v1","created_at":"2026-05-20T00:03:26Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.16857","created_at":"2026-05-20T00:03:26Z"},{"alias_kind":"pith_short_12","alias_value":"5OVVRCK3QAVK","created_at":"2026-05-20T00:03:26Z"},{"alias_kind":"pith_short_16","alias_value":"5OVVRCK3QAVKBZI6","created_at":"2026-05-20T00:03:26Z"},{"alias_kind":"pith_short_8","alias_value":"5OVVRCK3","created_at":"2026-05-20T00:03:26Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:5OVVRCK3QAVKBZI6OERYJRBOVH","target":"record","payload":{"canonical_record":{"source":{"id":"2605.16857","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2026-05-16T07:41:31Z","cross_cats_sorted":[],"title_canon_sha256":"3faf8d33f20e447ff83f46d9344b3ffead9f7bbd8baf6157d7d9f123318951b4","abstract_canon_sha256":"382ab370292499827e39db0e9bec48cbe2f9635e8d4c452a1c677b52a8f58039"},"schema_version":"1.0"},"canonical_sha256":"ebab58895b802aa0e51e712384c42ea9e1cea3f8ed467c73aad54eb7aa66cabf","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-20T00:03:26.572161Z","signature_b64":"e5T08F82lOQcsx7EobCUTrFFqE0xLOOzhgUS5O3xbxh/k0yZNozDrK+V/CRCC2cMNXonmfrXQ9yhCQgQ+nYvBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"ebab58895b802aa0e51e712384c42ea9e1cea3f8ed467c73aad54eb7aa66cabf","last_reissued_at":"2026-05-20T00:03:26.571243Z","signature_status":"signed_v1","first_computed_at":"2026-05-20T00:03:26.571243Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2605.16857","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":"2026-05-20T00:03:26Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"LpZA4Nuo+/X6HZmjYwvpO3WbQWr3Xdi/Q3cgwDzPu6Cmo4OkFuxt4YydPZNHhqhlxB15lKgN2wWS8LPCvD6KCg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-23T09:53:32.489272Z"},"content_sha256":"814e6f8efc45c75ff7fe29446a594b1dbe83ce5d0b32d2cc3c17fac170631fed","schema_version":"1.0","event_id":"sha256:814e6f8efc45c75ff7fe29446a594b1dbe83ce5d0b32d2cc3c17fac170631fed"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:5OVVRCK3QAVKBZI6OERYJRBOVH","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Learning to Learn from Multimodal Experience","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"Agents can substantially improve their performance on multimodal tasks by learning to dynamically construct and use memory from their experiences rather than depending on fixed designs.","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Bing Qin, Dandan Tu, Weixiang Zhao, Xingyu Sui, Yang Wu, Yanyan Zhao, Yongxin Tang","submitted_at":"2026-05-16T07:41:31Z","abstract_excerpt":"Experience-driven learning has emerged as a promising paradigm for enabling agents to improve from interaction trajectories by accumulating and reusing past experience. However, existing approaches are predominantly developed in textual settings and rely on manually designed memory schemas, limiting their applicability to multimodal environments. In real-world scenarios, experience is inherently multimodal, involving heterogeneous signals across perception, reasoning, and action, which makes effective memory design significantly more challenging. In particular, the optimal way to structure and"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"adaptive memory design substantially enhances agent performance and generalization across multimodal tasks","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"the optimal way to structure and utilize multimodal experience is highly task-dependent and evolves over time, rendering fixed memory designs insufficient","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"Agents learn to dynamically construct and organize memory from multimodal experiences, improving performance over static designs in task-dependent settings.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Agents can substantially improve their performance on multimodal tasks by learning to dynamically construct and use memory from their experiences rather than depending on fixed designs.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"9e63aaa21a61f2657d2c351f25e8df606a27830882965544f20fea6c3054b27e"},"source":{"id":"2605.16857","kind":"arxiv","version":1},"verdict":{"id":"a775926b-d789-4e03-b200-2cee017791d0","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-19T21:02:33.605757Z","strongest_claim":"adaptive memory design substantially enhances agent performance and generalization across multimodal tasks","one_line_summary":"Agents learn to dynamically construct and organize memory from multimodal experiences, improving performance over static designs in task-dependent settings.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"the optimal way to structure and utilize multimodal experience is highly task-dependent and evolves over time, rendering fixed memory designs insufficient","pith_extraction_headline":"Agents can substantially improve their performance on multimodal tasks by learning to dynamically construct and use memory from their experiences rather than depending on fixed designs."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2605.16857/integrity.json","findings":[],"available":true,"detectors_run":[{"name":"doi_title_agreement","ran_at":"2026-05-19T21:31:19.218421Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_compliance","ran_at":"2026-05-19T21:11:42.809886Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"claim_evidence","ran_at":"2026-05-19T18:41:56.308600Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"ai_meta_artifact","ran_at":"2026-05-19T18:33:26.383429Z","status":"skipped","version":"1.0.0","findings_count":0}],"snapshot_sha256":"42afaa69dbffab5f3aea720eb97f79c6e470436ac7cb7db908fcc130a75428c9"},"references":{"count":65,"sample":[{"doi":"","year":null,"title":"text“ (“str“, optional): guidance for the execution agent. - “images“ (optional): list of image refs with the same JSON shape as trajectory “ImageRef“: “{","work_id":"9d564306-32d8-4f3f-bd80-f3299e80d3b8","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":null,"title":"**examples** - **examples**: sampled retrieve trajectories, split into **FAILED TRAJECTORIES** and **SUCCESSFUL TRAJECTORIES**. - In each trajectory section below, every episode may include a text blo","work_id":"f7e821e4-3372-4663-9c5e-0018832bf35f","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":null,"title":"Need to use the score to analyze the performance and bottleneck of current memory structure","work_id":"9a4baa36-fba4-433b-b32b-928d36991592","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":null,"title":"Look at the provided improve_score (positive → improvement, negative → degradation) and the single suggestion_example that produced that score","work_id":"169524d6-58c1-4efe-888c-18981ae412e0","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":null,"title":"Step 2 — Inspect sampled trajectories and benchmark performance and decide which memories are useful","work_id":"281ff462-e029-46dd-a430-fb7eabb86d49","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":65,"snapshot_sha256":"38b6f1109c5d694d86063b8c6b4b4d7276dd9c85b0365cd201dc69064599ff20","internal_anchors":0},"formal_canon":{"evidence_count":1,"snapshot_sha256":"7f7e757cd3fdbe446de5d6e63f2f4a087bead21fdbe4ff008f2d1d1d1327c30e"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"verdict_id":"a775926b-d789-4e03-b200-2cee017791d0"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-20T00:03:26Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"rnmQ96ejbQc5+cMEESnA5efhhE9VY/AwAT4LCvXlUcN6lUzj1KNcYDT2lvwgRO5jXomz0tzIWs50AXE5KNpsBQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-23T09:53:32.490788Z"},"content_sha256":"e4b38d0ecf35b6ecac1b52d40a6317f0097b64981419fb5fa58461a03b17b47f","schema_version":"1.0","event_id":"sha256:e4b38d0ecf35b6ecac1b52d40a6317f0097b64981419fb5fa58461a03b17b47f"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/5OVVRCK3QAVKBZI6OERYJRBOVH/bundle.json","state_url":"https://pith.science/pith/5OVVRCK3QAVKBZI6OERYJRBOVH/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/5OVVRCK3QAVKBZI6OERYJRBOVH/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":"2026-05-23T09:53:32Z","links":{"resolver":"https://pith.science/pith/5OVVRCK3QAVKBZI6OERYJRBOVH","bundle":"https://pith.science/pith/5OVVRCK3QAVKBZI6OERYJRBOVH/bundle.json","state":"https://pith.science/pith/5OVVRCK3QAVKBZI6OERYJRBOVH/state.json","well_known_bundle":"https://pith.science/.well-known/pith/5OVVRCK3QAVKBZI6OERYJRBOVH/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:5OVVRCK3QAVKBZI6OERYJRBOVH","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":"382ab370292499827e39db0e9bec48cbe2f9635e8d4c452a1c677b52a8f58039","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2026-05-16T07:41:31Z","title_canon_sha256":"3faf8d33f20e447ff83f46d9344b3ffead9f7bbd8baf6157d7d9f123318951b4"},"schema_version":"1.0","source":{"id":"2605.16857","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.16857","created_at":"2026-05-20T00:03:26Z"},{"alias_kind":"arxiv_version","alias_value":"2605.16857v1","created_at":"2026-05-20T00:03:26Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.16857","created_at":"2026-05-20T00:03:26Z"},{"alias_kind":"pith_short_12","alias_value":"5OVVRCK3QAVK","created_at":"2026-05-20T00:03:26Z"},{"alias_kind":"pith_short_16","alias_value":"5OVVRCK3QAVKBZI6","created_at":"2026-05-20T00:03:26Z"},{"alias_kind":"pith_short_8","alias_value":"5OVVRCK3","created_at":"2026-05-20T00:03:26Z"}],"graph_snapshots":[{"event_id":"sha256:e4b38d0ecf35b6ecac1b52d40a6317f0097b64981419fb5fa58461a03b17b47f","target":"graph","created_at":"2026-05-20T00:03:26Z","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":"adaptive memory design substantially enhances agent performance and generalization across multimodal tasks"},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"the optimal way to structure and utilize multimodal experience is highly task-dependent and evolves over time, rendering fixed memory designs insufficient"},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"Agents learn to dynamically construct and organize memory from multimodal experiences, improving performance over static designs in task-dependent settings."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"Agents can substantially improve their performance on multimodal tasks by learning to dynamically construct and use memory from their experiences rather than depending on fixed designs."}],"snapshot_sha256":"9e63aaa21a61f2657d2c351f25e8df606a27830882965544f20fea6c3054b27e"},"formal_canon":{"evidence_count":1,"snapshot_sha256":"7f7e757cd3fdbe446de5d6e63f2f4a087bead21fdbe4ff008f2d1d1d1327c30e"},"integrity":{"available":true,"clean":true,"detectors_run":[{"findings_count":0,"name":"doi_title_agreement","ran_at":"2026-05-19T21:31:19.218421Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"doi_compliance","ran_at":"2026-05-19T21:11:42.809886Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"claim_evidence","ran_at":"2026-05-19T18:41:56.308600Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"ai_meta_artifact","ran_at":"2026-05-19T18:33:26.383429Z","status":"skipped","version":"1.0.0"}],"endpoint":"/pith/2605.16857/integrity.json","findings":[],"snapshot_sha256":"42afaa69dbffab5f3aea720eb97f79c6e470436ac7cb7db908fcc130a75428c9","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Experience-driven learning has emerged as a promising paradigm for enabling agents to improve from interaction trajectories by accumulating and reusing past experience. However, existing approaches are predominantly developed in textual settings and rely on manually designed memory schemas, limiting their applicability to multimodal environments. In real-world scenarios, experience is inherently multimodal, involving heterogeneous signals across perception, reasoning, and action, which makes effective memory design significantly more challenging. In particular, the optimal way to structure and","authors_text":"Bing Qin, Dandan Tu, Weixiang Zhao, Xingyu Sui, Yang Wu, Yanyan Zhao, Yongxin Tang","cross_cats":[],"headline":"Agents can substantially improve their performance on multimodal tasks by learning to dynamically construct and use memory from their experiences rather than depending on fixed designs.","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2026-05-16T07:41:31Z","title":"Learning to Learn from Multimodal Experience"},"references":{"count":65,"internal_anchors":0,"resolved_work":65,"sample":[{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":1,"title":"text“ (“str“, optional): guidance for the execution agent. - “images“ (optional): list of image refs with the same JSON shape as trajectory “ImageRef“: “{","work_id":"9d564306-32d8-4f3f-bd80-f3299e80d3b8","year":null},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":2,"title":"**examples** - **examples**: sampled retrieve trajectories, split into **FAILED TRAJECTORIES** and **SUCCESSFUL TRAJECTORIES**. - In each trajectory section below, every episode may include a text blo","work_id":"f7e821e4-3372-4663-9c5e-0018832bf35f","year":null},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":3,"title":"Need to use the score to analyze the performance and bottleneck of current memory structure","work_id":"9a4baa36-fba4-433b-b32b-928d36991592","year":null},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":4,"title":"Look at the provided improve_score (positive → improvement, negative → degradation) and the single suggestion_example that produced that score","work_id":"169524d6-58c1-4efe-888c-18981ae412e0","year":null},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":5,"title":"Step 2 — Inspect sampled trajectories and benchmark performance and decide which memories are useful","work_id":"281ff462-e029-46dd-a430-fb7eabb86d49","year":null}],"snapshot_sha256":"38b6f1109c5d694d86063b8c6b4b4d7276dd9c85b0365cd201dc69064599ff20"},"source":{"id":"2605.16857","kind":"arxiv","version":1},"verdict":{"created_at":"2026-05-19T21:02:33.605757Z","id":"a775926b-d789-4e03-b200-2cee017791d0","model_set":{"reader":"grok-4.3"},"one_line_summary":"Agents learn to dynamically construct and organize memory from multimodal experiences, improving performance over static designs in task-dependent settings.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"Agents can substantially improve their performance on multimodal tasks by learning to dynamically construct and use memory from their experiences rather than depending on fixed designs.","strongest_claim":"adaptive memory design substantially enhances agent performance and generalization across multimodal tasks","weakest_assumption":"the optimal way to structure and utilize multimodal experience is highly task-dependent and evolves over time, rendering fixed memory designs insufficient"}},"verdict_id":"a775926b-d789-4e03-b200-2cee017791d0"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:814e6f8efc45c75ff7fe29446a594b1dbe83ce5d0b32d2cc3c17fac170631fed","target":"record","created_at":"2026-05-20T00:03:26Z","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":"382ab370292499827e39db0e9bec48cbe2f9635e8d4c452a1c677b52a8f58039","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2026-05-16T07:41:31Z","title_canon_sha256":"3faf8d33f20e447ff83f46d9344b3ffead9f7bbd8baf6157d7d9f123318951b4"},"schema_version":"1.0","source":{"id":"2605.16857","kind":"arxiv","version":1}},"canonical_sha256":"ebab58895b802aa0e51e712384c42ea9e1cea3f8ed467c73aad54eb7aa66cabf","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"ebab58895b802aa0e51e712384c42ea9e1cea3f8ed467c73aad54eb7aa66cabf","first_computed_at":"2026-05-20T00:03:26.571243Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-20T00:03:26.571243Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"e5T08F82lOQcsx7EobCUTrFFqE0xLOOzhgUS5O3xbxh/k0yZNozDrK+V/CRCC2cMNXonmfrXQ9yhCQgQ+nYvBQ==","signature_status":"signed_v1","signed_at":"2026-05-20T00:03:26.572161Z","signed_message":"canonical_sha256_bytes"},"source_id":"2605.16857","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:814e6f8efc45c75ff7fe29446a594b1dbe83ce5d0b32d2cc3c17fac170631fed","sha256:e4b38d0ecf35b6ecac1b52d40a6317f0097b64981419fb5fa58461a03b17b47f"],"state_sha256":"67e79e9ba38827a526b7bdd3db056c799932abbf5d88a1fe7aa157a8b3b06db6"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"g/Jiyh1Sbs8DzYM2Rn59ZCyk20Zm1q7NWeC0e30qJVct/Vx76eYkuxMzI6onpvqwauddG9B0v84Ojl6+fKiKBw==","signed_message":"bundle_sha256_bytes","signed_at":"2026-05-23T09:53:32.496775Z","bundle_sha256":"6a9346a34398a8f5a7d97659798c6f4c94bc927135ea1665affcbc0ddbac7760"}}