{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:AZUJNQMBN7Y62AOZHD3PNPWCZ3","short_pith_number":"pith:AZUJNQMB","canonical_record":{"source":{"id":"2605.13010","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2026-05-13T05:02:47Z","cross_cats_sorted":["cs.AI","cs.SY","eess.SY","math.OC"],"title_canon_sha256":"0569d971af108bba9a317fbc5a7809cded99d0557026a7fbe21283e903c98b29","abstract_canon_sha256":"8feac85893658852073a0a19728f86246c74cfc295a8e6920c3baab05ccf80b6"},"schema_version":"1.0"},"canonical_sha256":"066896c1816ff1ed01d938f6f6bec2ceff3bfc7406b5e629b82a7a6f7746799b","source":{"kind":"arxiv","id":"2605.13010","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.13010","created_at":"2026-05-18T03:09:00Z"},{"alias_kind":"arxiv_version","alias_value":"2605.13010v1","created_at":"2026-05-18T03:09:00Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.13010","created_at":"2026-05-18T03:09:00Z"},{"alias_kind":"pith_short_12","alias_value":"AZUJNQMBN7Y6","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_16","alias_value":"AZUJNQMBN7Y62AOZ","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_8","alias_value":"AZUJNQMB","created_at":"2026-05-18T12:33:37Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:AZUJNQMBN7Y62AOZHD3PNPWCZ3","target":"record","payload":{"canonical_record":{"source":{"id":"2605.13010","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2026-05-13T05:02:47Z","cross_cats_sorted":["cs.AI","cs.SY","eess.SY","math.OC"],"title_canon_sha256":"0569d971af108bba9a317fbc5a7809cded99d0557026a7fbe21283e903c98b29","abstract_canon_sha256":"8feac85893658852073a0a19728f86246c74cfc295a8e6920c3baab05ccf80b6"},"schema_version":"1.0"},"canonical_sha256":"066896c1816ff1ed01d938f6f6bec2ceff3bfc7406b5e629b82a7a6f7746799b","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-18T03:09:00.262551Z","signature_b64":"GgL9DRoKqTx+iaJhDuC5jbh4rWMZMqrAPN9TnE23SSoEyU4WhPOAmk2FJV3uX6AJdUQVfZ9vCVtYgMTVDvGtAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"066896c1816ff1ed01d938f6f6bec2ceff3bfc7406b5e629b82a7a6f7746799b","last_reissued_at":"2026-05-18T03:09:00.261970Z","signature_status":"signed_v1","first_computed_at":"2026-05-18T03:09:00.261970Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2605.13010","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":"2026-05-18T03:09:00Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"M8Yer72Bgck9NGAg5w/j8Z+IDFq7ePb21Xk1JGDxcqLbyNw23I67Gmj09LmJ0htgrqk4gAFtNk7B3hWYKB1vCQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-07T01:48:55.198706Z"},"content_sha256":"eb4586f17a792c34422c42247d517ab55e80476f1e5348bf5e54ce0ab6d63491","schema_version":"1.0","event_id":"sha256:eb4586f17a792c34422c42247d517ab55e80476f1e5348bf5e54ce0ab6d63491"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:AZUJNQMBN7Y62AOZHD3PNPWCZ3","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Amortized Guidance for Image Inpainting with Pretrained Diffusion Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"A small guidance module trained once offline guides pretrained diffusion models for faster inpainting without per-image optimization.","cross_cats":["cs.AI","cs.SY","eess.SY","math.OC"],"primary_cat":"cs.CV","authors_text":"Xun Yu Zhou, Yilie Huang","submitted_at":"2026-05-13T05:02:47Z","abstract_excerpt":"We study image inpainting with generative diffusion models. Existing methods typically either train dedicated task-specific models, or adapt a pretrained diffusion model separately for each masked image at deployment. We introduce a middle-ground model, termed Amortized Inpainting with Diffusion (AID), which keeps a pretrained diffusion backbone fixed, trains a small reusable guidance module offline, and then reuses it across masked images without per-instance optimization. We formulate it as a deterministic guidance problem with a supervised terminal objective. To make this problem learnable "},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"We introduce a middle-ground model, termed Amortized Inpainting with Diffusion (AID), which keeps a pretrained diffusion backbone fixed, trains a small reusable guidance module offline, and then reuses it across masked images without per-instance optimization. We derive an auxiliary Gaussian formulation and prove that solving this randomized problem recovers the optimal deterministic guidance field.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"The auxiliary Gaussian formulation and the proof that its solution recovers the optimal deterministic guidance field hold in high-dimensional image spaces; this bridge is required for the continuous-time actor-critic algorithm to be valid.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"AID amortizes guidance for diffusion inpainting by training a reusable module via an auxiliary Gaussian formulation and continuous-time actor-critic algorithm, improving quality-speed trade-off with under 1% overhead.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"A small guidance module trained once offline guides pretrained diffusion models for faster inpainting without per-image optimization.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"22356cd534bab169d2925c79d5514e4cf2dde66b3525a58f097a0458b52bb04c"},"source":{"id":"2605.13010","kind":"arxiv","version":1},"verdict":{"id":"324df07b-1299-42bd-ad94-871171095c49","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-14T19:29:34.243031Z","strongest_claim":"We introduce a middle-ground model, termed Amortized Inpainting with Diffusion (AID), which keeps a pretrained diffusion backbone fixed, trains a small reusable guidance module offline, and then reuses it across masked images without per-instance optimization. We derive an auxiliary Gaussian formulation and prove that solving this randomized problem recovers the optimal deterministic guidance field.","one_line_summary":"AID amortizes guidance for diffusion inpainting by training a reusable module via an auxiliary Gaussian formulation and continuous-time actor-critic algorithm, improving quality-speed trade-off with under 1% overhead.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"The auxiliary Gaussian formulation and the proof that its solution recovers the optimal deterministic guidance field hold in high-dimensional image spaces; this bridge is required for the continuous-time actor-critic algorithm to be valid.","pith_extraction_headline":"A small guidance module trained once offline guides pretrained diffusion models for faster inpainting without per-image optimization."},"references":{"count":41,"sample":[{"doi":"","year":2020,"title":"StarGAN v2: Diverse image synthesis for multiple domains","work_id":"530057f9-24bf-49a1-a0c2-4012cebbdcce","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2022,"title":"Diffusion Posterior Sampling for General Noisy Inverse Problems","work_id":"083ab9fb-05f0-41e1-9628-982017eac344","ref_index":2,"cited_arxiv_id":"2209.14687","is_internal_anchor":true},{"doi":"","year":2022,"title":"Improving diffusion models for inverse problems using manifold constraints.Advances in Neural Information Processing Systems, 35: 25683–25696, 2022","work_id":"44074e66-f0d4-4680-a8f7-7526c46f8f18","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2024,"title":"LatentPaint: Image inpainting in latent space with diffusion models","work_id":"b35c9c5e-1d1b-450c-a5f5-a56158b964f0","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2009,"title":"Imagenet: A large-scale hierarchical image database","work_id":"cc291e4b-478b-4e79-ab37-d782c8e1888e","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":41,"snapshot_sha256":"0c7633a76da8fe4b2efbe70e5e69296bcf8de800a585f8a74fe29129797d72c7","internal_anchors":6},"formal_canon":{"evidence_count":2,"snapshot_sha256":"5c331e3ed8f6ba1a5edf1ea8d443ac659b26525ae1e74683a7089dc0ab235af1"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"verdict_id":"324df07b-1299-42bd-ad94-871171095c49"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-18T03:09:00Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"887Evc0uD/QXGRXYGTjso1N2v+B9belQWr2s09NFeKfVgnhEZOzr88rh1ej6+VZbGeX6SZlFbtIqJx9pMgdOBg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-07T01:48:55.199979Z"},"content_sha256":"9f024cc764ebe04c3ca67a8fc70ecc623a7fc16f89016a1b3ab06aeae3b956e7","schema_version":"1.0","event_id":"sha256:9f024cc764ebe04c3ca67a8fc70ecc623a7fc16f89016a1b3ab06aeae3b956e7"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/AZUJNQMBN7Y62AOZHD3PNPWCZ3/bundle.json","state_url":"https://pith.science/pith/AZUJNQMBN7Y62AOZHD3PNPWCZ3/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/AZUJNQMBN7Y62AOZHD3PNPWCZ3/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":"2026-06-07T01:48:55Z","links":{"resolver":"https://pith.science/pith/AZUJNQMBN7Y62AOZHD3PNPWCZ3","bundle":"https://pith.science/pith/AZUJNQMBN7Y62AOZHD3PNPWCZ3/bundle.json","state":"https://pith.science/pith/AZUJNQMBN7Y62AOZHD3PNPWCZ3/state.json","well_known_bundle":"https://pith.science/.well-known/pith/AZUJNQMBN7Y62AOZHD3PNPWCZ3/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:AZUJNQMBN7Y62AOZHD3PNPWCZ3","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":"8feac85893658852073a0a19728f86246c74cfc295a8e6920c3baab05ccf80b6","cross_cats_sorted":["cs.AI","cs.SY","eess.SY","math.OC"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2026-05-13T05:02:47Z","title_canon_sha256":"0569d971af108bba9a317fbc5a7809cded99d0557026a7fbe21283e903c98b29"},"schema_version":"1.0","source":{"id":"2605.13010","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.13010","created_at":"2026-05-18T03:09:00Z"},{"alias_kind":"arxiv_version","alias_value":"2605.13010v1","created_at":"2026-05-18T03:09:00Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.13010","created_at":"2026-05-18T03:09:00Z"},{"alias_kind":"pith_short_12","alias_value":"AZUJNQMBN7Y6","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_16","alias_value":"AZUJNQMBN7Y62AOZ","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_8","alias_value":"AZUJNQMB","created_at":"2026-05-18T12:33:37Z"}],"graph_snapshots":[{"event_id":"sha256:9f024cc764ebe04c3ca67a8fc70ecc623a7fc16f89016a1b3ab06aeae3b956e7","target":"graph","created_at":"2026-05-18T03:09:00Z","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":"We introduce a middle-ground model, termed Amortized Inpainting with Diffusion (AID), which keeps a pretrained diffusion backbone fixed, trains a small reusable guidance module offline, and then reuses it across masked images without per-instance optimization. We derive an auxiliary Gaussian formulation and prove that solving this randomized problem recovers the optimal deterministic guidance field."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"The auxiliary Gaussian formulation and the proof that its solution recovers the optimal deterministic guidance field hold in high-dimensional image spaces; this bridge is required for the continuous-time actor-critic algorithm to be valid."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"AID amortizes guidance for diffusion inpainting by training a reusable module via an auxiliary Gaussian formulation and continuous-time actor-critic algorithm, improving quality-speed trade-off with under 1% overhead."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"A small guidance module trained once offline guides pretrained diffusion models for faster inpainting without per-image optimization."}],"snapshot_sha256":"22356cd534bab169d2925c79d5514e4cf2dde66b3525a58f097a0458b52bb04c"},"formal_canon":{"evidence_count":2,"snapshot_sha256":"5c331e3ed8f6ba1a5edf1ea8d443ac659b26525ae1e74683a7089dc0ab235af1"},"paper":{"abstract_excerpt":"We study image inpainting with generative diffusion models. Existing methods typically either train dedicated task-specific models, or adapt a pretrained diffusion model separately for each masked image at deployment. We introduce a middle-ground model, termed Amortized Inpainting with Diffusion (AID), which keeps a pretrained diffusion backbone fixed, trains a small reusable guidance module offline, and then reuses it across masked images without per-instance optimization. We formulate it as a deterministic guidance problem with a supervised terminal objective. To make this problem learnable ","authors_text":"Xun Yu Zhou, Yilie Huang","cross_cats":["cs.AI","cs.SY","eess.SY","math.OC"],"headline":"A small guidance module trained once offline guides pretrained diffusion models for faster inpainting without per-image optimization.","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2026-05-13T05:02:47Z","title":"Amortized Guidance for Image Inpainting with Pretrained Diffusion Models"},"references":{"count":41,"internal_anchors":6,"resolved_work":41,"sample":[{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":1,"title":"StarGAN v2: Diverse image synthesis for multiple domains","work_id":"530057f9-24bf-49a1-a0c2-4012cebbdcce","year":2020},{"cited_arxiv_id":"2209.14687","doi":"","is_internal_anchor":true,"ref_index":2,"title":"Diffusion Posterior Sampling for General Noisy Inverse Problems","work_id":"083ab9fb-05f0-41e1-9628-982017eac344","year":2022},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":3,"title":"Improving diffusion models for inverse problems using manifold constraints.Advances in Neural Information Processing Systems, 35: 25683–25696, 2022","work_id":"44074e66-f0d4-4680-a8f7-7526c46f8f18","year":2022},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":4,"title":"LatentPaint: Image inpainting in latent space with diffusion models","work_id":"b35c9c5e-1d1b-450c-a5f5-a56158b964f0","year":2024},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":5,"title":"Imagenet: A large-scale hierarchical image database","work_id":"cc291e4b-478b-4e79-ab37-d782c8e1888e","year":2009}],"snapshot_sha256":"0c7633a76da8fe4b2efbe70e5e69296bcf8de800a585f8a74fe29129797d72c7"},"source":{"id":"2605.13010","kind":"arxiv","version":1},"verdict":{"created_at":"2026-05-14T19:29:34.243031Z","id":"324df07b-1299-42bd-ad94-871171095c49","model_set":{"reader":"grok-4.3"},"one_line_summary":"AID amortizes guidance for diffusion inpainting by training a reusable module via an auxiliary Gaussian formulation and continuous-time actor-critic algorithm, improving quality-speed trade-off with under 1% overhead.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"A small guidance module trained once offline guides pretrained diffusion models for faster inpainting without per-image optimization.","strongest_claim":"We introduce a middle-ground model, termed Amortized Inpainting with Diffusion (AID), which keeps a pretrained diffusion backbone fixed, trains a small reusable guidance module offline, and then reuses it across masked images without per-instance optimization. We derive an auxiliary Gaussian formulation and prove that solving this randomized problem recovers the optimal deterministic guidance field.","weakest_assumption":"The auxiliary Gaussian formulation and the proof that its solution recovers the optimal deterministic guidance field hold in high-dimensional image spaces; this bridge is required for the continuous-time actor-critic algorithm to be valid."}},"verdict_id":"324df07b-1299-42bd-ad94-871171095c49"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:eb4586f17a792c34422c42247d517ab55e80476f1e5348bf5e54ce0ab6d63491","target":"record","created_at":"2026-05-18T03:09:00Z","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":"8feac85893658852073a0a19728f86246c74cfc295a8e6920c3baab05ccf80b6","cross_cats_sorted":["cs.AI","cs.SY","eess.SY","math.OC"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2026-05-13T05:02:47Z","title_canon_sha256":"0569d971af108bba9a317fbc5a7809cded99d0557026a7fbe21283e903c98b29"},"schema_version":"1.0","source":{"id":"2605.13010","kind":"arxiv","version":1}},"canonical_sha256":"066896c1816ff1ed01d938f6f6bec2ceff3bfc7406b5e629b82a7a6f7746799b","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"066896c1816ff1ed01d938f6f6bec2ceff3bfc7406b5e629b82a7a6f7746799b","first_computed_at":"2026-05-18T03:09:00.261970Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-18T03:09:00.261970Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"GgL9DRoKqTx+iaJhDuC5jbh4rWMZMqrAPN9TnE23SSoEyU4WhPOAmk2FJV3uX6AJdUQVfZ9vCVtYgMTVDvGtAw==","signature_status":"signed_v1","signed_at":"2026-05-18T03:09:00.262551Z","signed_message":"canonical_sha256_bytes"},"source_id":"2605.13010","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:eb4586f17a792c34422c42247d517ab55e80476f1e5348bf5e54ce0ab6d63491","sha256:9f024cc764ebe04c3ca67a8fc70ecc623a7fc16f89016a1b3ab06aeae3b956e7"],"state_sha256":"3b84fc82e37f96d498236b9ffa9e7f96ddaa2fed3b1b2aab1f37dcd36047ef28"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"1lCkly8Twk5/KA0dAayikhXGCvffhdXCWLWkY4H9psim4MPu6NGDby926H9bAw+6nov5CBpGTXn/QFIezrOxCw==","signed_message":"bundle_sha256_bytes","signed_at":"2026-06-07T01:48:55.205632Z","bundle_sha256":"0ce14810c954d3061d513e4c9b6630764bcc2016bfab197745b8b86ce744367c"}}