{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:UP7KD3NGBK45HYPYKDVXMI33OY","short_pith_number":"pith:UP7KD3NG","schema_version":"1.0","canonical_sha256":"a3fea1eda60ab9d3e1f850eb76237b763e6e104d3e826e66ed69c31e6af9be32","source":{"kind":"arxiv","id":"2506.20737","version":1},"attestation_state":"computed","paper":{"title":"MAGPIE: A dataset for Multi-AGent contextual PrIvacy Evaluation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.AI","authors_text":"Alon Albalak, Gurusha Juneja, Wenyue Hua, William Yang Wang","submitted_at":"2025-06-25T18:04:25Z","abstract_excerpt":"The proliferation of LLM-based agents has led to increasing deployment of inter-agent collaboration for tasks like scheduling, negotiation, resource allocation etc. In such systems, privacy is critical, as agents often access proprietary tools and domain-specific databases requiring strict confidentiality. This paper examines whether LLM-based agents demonstrate an understanding of contextual privacy. And, if instructed, do these systems preserve inference time user privacy in non-adversarial multi-turn conversation. Existing benchmarks to evaluate contextual privacy in LLM-agents primarily as"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2506.20737","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2025-06-25T18:04:25Z","cross_cats_sorted":["cs.CL"],"title_canon_sha256":"6d2f8fb2e7143d6c13fe00aa06da7c676e7548dd8655def377ea28c6448e9e56","abstract_canon_sha256":"26f5ec9dffe086b96a343873db59d38a9bf15ae2b95d116a90ca06f0e5f1f6dd"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:27:33.234975Z","signature_b64":"oetVU1YfEIrbHN5B1mLpbk1v72ATZtoYtlqr7Q34TeDTVkQOPGJEsQ06QhEU8zqdEChIcHyN4HaavEaFQpGvDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"a3fea1eda60ab9d3e1f850eb76237b763e6e104d3e826e66ed69c31e6af9be32","last_reissued_at":"2026-07-05T11:27:33.234481Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:27:33.234481Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"MAGPIE: A dataset for Multi-AGent contextual PrIvacy Evaluation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.AI","authors_text":"Alon Albalak, Gurusha Juneja, Wenyue Hua, William Yang Wang","submitted_at":"2025-06-25T18:04:25Z","abstract_excerpt":"The proliferation of LLM-based agents has led to increasing deployment of inter-agent collaboration for tasks like scheduling, negotiation, resource allocation etc. In such systems, privacy is critical, as agents often access proprietary tools and domain-specific databases requiring strict confidentiality. This paper examines whether LLM-based agents demonstrate an understanding of contextual privacy. And, if instructed, do these systems preserve inference time user privacy in non-adversarial multi-turn conversation. Existing benchmarks to evaluate contextual privacy in LLM-agents primarily as"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.20737","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/2506.20737/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"},"aliases":[{"alias_kind":"arxiv","alias_value":"2506.20737","created_at":"2026-07-05T11:27:33.234542+00:00"},{"alias_kind":"arxiv_version","alias_value":"2506.20737v1","created_at":"2026-07-05T11:27:33.234542+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.20737","created_at":"2026-07-05T11:27:33.234542+00:00"},{"alias_kind":"pith_short_12","alias_value":"UP7KD3NGBK45","created_at":"2026-07-05T11:27:33.234542+00:00"},{"alias_kind":"pith_short_16","alias_value":"UP7KD3NGBK45HYPY","created_at":"2026-07-05T11:27:33.234542+00:00"},{"alias_kind":"pith_short_8","alias_value":"UP7KD3NG","created_at":"2026-07-05T11:27:33.234542+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2603.09002","citing_title":"Security Considerations for Multi-agent Systems","ref_index":168,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/UP7KD3NGBK45HYPYKDVXMI33OY","json":"https://pith.science/pith/UP7KD3NGBK45HYPYKDVXMI33OY.json","graph_json":"https://pith.science/api/pith-number/UP7KD3NGBK45HYPYKDVXMI33OY/graph.json","events_json":"https://pith.science/api/pith-number/UP7KD3NGBK45HYPYKDVXMI33OY/events.json","paper":"https://pith.science/paper/UP7KD3NG"},"agent_actions":{"view_html":"https://pith.science/pith/UP7KD3NGBK45HYPYKDVXMI33OY","download_json":"https://pith.science/pith/UP7KD3NGBK45HYPYKDVXMI33OY.json","view_paper":"https://pith.science/paper/UP7KD3NG","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2506.20737&json=true","fetch_graph":"https://pith.science/api/pith-number/UP7KD3NGBK45HYPYKDVXMI33OY/graph.json","fetch_events":"https://pith.science/api/pith-number/UP7KD3NGBK45HYPYKDVXMI33OY/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/UP7KD3NGBK45HYPYKDVXMI33OY/action/timestamp_anchor","attest_storage":"https://pith.science/pith/UP7KD3NGBK45HYPYKDVXMI33OY/action/storage_attestation","attest_author":"https://pith.science/pith/UP7KD3NGBK45HYPYKDVXMI33OY/action/author_attestation","sign_citation":"https://pith.science/pith/UP7KD3NGBK45HYPYKDVXMI33OY/action/citation_signature","submit_replication":"https://pith.science/pith/UP7KD3NGBK45HYPYKDVXMI33OY/action/replication_record"}},"created_at":"2026-07-05T11:27:33.234542+00:00","updated_at":"2026-07-05T11:27:33.234542+00:00"}