{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:5AMFRTNRNFOJWSOVS55K56ZZIW","short_pith_number":"pith:5AMFRTNR","canonical_record":{"source":{"id":"2506.09600","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.MA","submitted_at":"2025-06-11T10:59:47Z","cross_cats_sorted":["cs.AI","cs.CL","cs.CR"],"title_canon_sha256":"f7a628162d57834b9847718d6c863ece2094ff655501af4de0998a9054585fb5","abstract_canon_sha256":"7bb49934d91d8bf0a74f9edf922891f765df8579f84a2522d0e3b2e1080fe82a"},"schema_version":"1.0"},"canonical_sha256":"e81858cdb1695c9b49d5977aaefb3945819ba852f8367f4af02c515170993d81","source":{"kind":"arxiv","id":"2506.09600","version":3},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2506.09600","created_at":"2026-07-05T11:57:57Z"},{"alias_kind":"arxiv_version","alias_value":"2506.09600v3","created_at":"2026-07-05T11:57:57Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.09600","created_at":"2026-07-05T11:57:57Z"},{"alias_kind":"pith_short_12","alias_value":"5AMFRTNRNFOJ","created_at":"2026-07-05T11:57:57Z"},{"alias_kind":"pith_short_16","alias_value":"5AMFRTNRNFOJWSOV","created_at":"2026-07-05T11:57:57Z"},{"alias_kind":"pith_short_8","alias_value":"5AMFRTNR","created_at":"2026-07-05T11:57:57Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:5AMFRTNRNFOJWSOVS55K56ZZIW","target":"record","payload":{"canonical_record":{"source":{"id":"2506.09600","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.MA","submitted_at":"2025-06-11T10:59:47Z","cross_cats_sorted":["cs.AI","cs.CL","cs.CR"],"title_canon_sha256":"f7a628162d57834b9847718d6c863ece2094ff655501af4de0998a9054585fb5","abstract_canon_sha256":"7bb49934d91d8bf0a74f9edf922891f765df8579f84a2522d0e3b2e1080fe82a"},"schema_version":"1.0"},"canonical_sha256":"e81858cdb1695c9b49d5977aaefb3945819ba852f8367f4af02c515170993d81","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:57:57.118402Z","signature_b64":"T4Hn2L+0ciUAHl3W6S9fmYZ0/90Hm4CnMo5G2DyB2DpzQFgNGuaQhF3KTqpULxdOqfOy7+GJG1RxQdo3g2doCg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e81858cdb1695c9b49d5977aaefb3945819ba852f8367f4af02c515170993d81","last_reissued_at":"2026-07-05T11:57:57.117720Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:57:57.117720Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2506.09600","source_version":3,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T11:57:57Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"O5XAE/vzqXbaFUV1xcEPSYOXwDlstMwfO7JfRzq0XgPbMVbqrbPr66nZ1G9hx0PE8wyM+yRBtkWVgwkXxvnZAQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-09T01:13:51.025993Z"},"content_sha256":"cc5b35a0d2a0a215648794a3f1c49c8cdc0db7b94338aaf443dd860ea56d0086","schema_version":"1.0","event_id":"sha256:cc5b35a0d2a0a215648794a3f1c49c8cdc0db7b94338aaf443dd860ea56d0086"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:5AMFRTNRNFOJWSOVS55K56ZZIW","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Effective Red-Teaming of Policy-Adherent Agents","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CL","cs.CR"],"primary_cat":"cs.MA","authors_text":"Ateret Anaby-Tavor, George Kour, Guy Uziel, Itay Nakash, Koren Lazar, Matan Vetzler","submitted_at":"2025-06-11T10:59:47Z","abstract_excerpt":"Task-oriented LLM-based agents are increasingly used in domains with strict policies, such as refund eligibility or cancellation rules. The challenge lies in ensuring that the agent consistently adheres to these rules and policies, appropriately refusing any request that would violate them, while still maintaining a helpful and natural interaction. This calls for the development of tailored design and evaluation methodologies to ensure agent resilience against malicious user behavior. We propose a novel threat model that focuses on adversarial users aiming to exploit policy-adherent agents for"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.09600","kind":"arxiv","version":3},"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/2506.09600/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"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T11:57:57Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"b77T7SzCiBrFbrJgPjNlrR6I4eUiRiGrrQJJNcHKkaNQy4CwUImvGdoS8qhn60aC3JrjL3hZ7kegWdmooiMDCA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-09T01:13:51.026554Z"},"content_sha256":"a583dec31093dcd9d3e42794f5e9363caec439c3981106510423289090ddff32","schema_version":"1.0","event_id":"sha256:a583dec31093dcd9d3e42794f5e9363caec439c3981106510423289090ddff32"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/5AMFRTNRNFOJWSOVS55K56ZZIW/bundle.json","state_url":"https://pith.science/pith/5AMFRTNRNFOJWSOVS55K56ZZIW/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/5AMFRTNRNFOJWSOVS55K56ZZIW/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":"2026-08-09T01:13:51Z","links":{"resolver":"https://pith.science/pith/5AMFRTNRNFOJWSOVS55K56ZZIW","bundle":"https://pith.science/pith/5AMFRTNRNFOJWSOVS55K56ZZIW/bundle.json","state":"https://pith.science/pith/5AMFRTNRNFOJWSOVS55K56ZZIW/state.json","well_known_bundle":"https://pith.science/.well-known/pith/5AMFRTNRNFOJWSOVS55K56ZZIW/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:5AMFRTNRNFOJWSOVS55K56ZZIW","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":"7bb49934d91d8bf0a74f9edf922891f765df8579f84a2522d0e3b2e1080fe82a","cross_cats_sorted":["cs.AI","cs.CL","cs.CR"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.MA","submitted_at":"2025-06-11T10:59:47Z","title_canon_sha256":"f7a628162d57834b9847718d6c863ece2094ff655501af4de0998a9054585fb5"},"schema_version":"1.0","source":{"id":"2506.09600","kind":"arxiv","version":3}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2506.09600","created_at":"2026-07-05T11:57:57Z"},{"alias_kind":"arxiv_version","alias_value":"2506.09600v3","created_at":"2026-07-05T11:57:57Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.09600","created_at":"2026-07-05T11:57:57Z"},{"alias_kind":"pith_short_12","alias_value":"5AMFRTNRNFOJ","created_at":"2026-07-05T11:57:57Z"},{"alias_kind":"pith_short_16","alias_value":"5AMFRTNRNFOJWSOV","created_at":"2026-07-05T11:57:57Z"},{"alias_kind":"pith_short_8","alias_value":"5AMFRTNR","created_at":"2026-07-05T11:57:57Z"}],"graph_snapshots":[{"event_id":"sha256:a583dec31093dcd9d3e42794f5e9363caec439c3981106510423289090ddff32","target":"graph","created_at":"2026-07-05T11:57:57Z","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":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2506.09600/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Task-oriented LLM-based agents are increasingly used in domains with strict policies, such as refund eligibility or cancellation rules. The challenge lies in ensuring that the agent consistently adheres to these rules and policies, appropriately refusing any request that would violate them, while still maintaining a helpful and natural interaction. This calls for the development of tailored design and evaluation methodologies to ensure agent resilience against malicious user behavior. We propose a novel threat model that focuses on adversarial users aiming to exploit policy-adherent agents for","authors_text":"Ateret Anaby-Tavor, George Kour, Guy Uziel, Itay Nakash, Koren Lazar, Matan Vetzler","cross_cats":["cs.AI","cs.CL","cs.CR"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.MA","submitted_at":"2025-06-11T10:59:47Z","title":"Effective Red-Teaming of Policy-Adherent Agents"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.09600","kind":"arxiv","version":3},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:cc5b35a0d2a0a215648794a3f1c49c8cdc0db7b94338aaf443dd860ea56d0086","target":"record","created_at":"2026-07-05T11:57:57Z","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":"7bb49934d91d8bf0a74f9edf922891f765df8579f84a2522d0e3b2e1080fe82a","cross_cats_sorted":["cs.AI","cs.CL","cs.CR"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.MA","submitted_at":"2025-06-11T10:59:47Z","title_canon_sha256":"f7a628162d57834b9847718d6c863ece2094ff655501af4de0998a9054585fb5"},"schema_version":"1.0","source":{"id":"2506.09600","kind":"arxiv","version":3}},"canonical_sha256":"e81858cdb1695c9b49d5977aaefb3945819ba852f8367f4af02c515170993d81","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"e81858cdb1695c9b49d5977aaefb3945819ba852f8367f4af02c515170993d81","first_computed_at":"2026-07-05T11:57:57.117720Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:57:57.117720Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"T4Hn2L+0ciUAHl3W6S9fmYZ0/90Hm4CnMo5G2DyB2DpzQFgNGuaQhF3KTqpULxdOqfOy7+GJG1RxQdo3g2doCg==","signature_status":"signed_v1","signed_at":"2026-07-05T11:57:57.118402Z","signed_message":"canonical_sha256_bytes"},"source_id":"2506.09600","source_kind":"arxiv","source_version":3}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:cc5b35a0d2a0a215648794a3f1c49c8cdc0db7b94338aaf443dd860ea56d0086","sha256:a583dec31093dcd9d3e42794f5e9363caec439c3981106510423289090ddff32"],"state_sha256":"18aac30fae91772471afee35c774050ec1bc063273e65739ddc6182ae805a4f2"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"q3iJgxzxi394O5wAyxxL8ne4cq7L236fbhBc/dV+Lf2prHHBkX2XpPETnvIn6z9Ize+XyvW+XFVPRB7Zdwf8BQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-09T01:13:51.030052Z","bundle_sha256":"d95316d57cb01345bc024fe9900d7cb24818f82174fd57fa0c14b5c312926956"}}