{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:PMJ2YIVLC5XEGJRWXGHLMIGMJF","short_pith_number":"pith:PMJ2YIVL","schema_version":"1.0","canonical_sha256":"7b13ac22ab176e432636b98eb620cc4977e5a0990019a7cf5c80d10b580e2824","source":{"kind":"arxiv","id":"2607.18063","version":1},"attestation_state":"computed","paper":{"title":"Adaptive Adversaries: A Multi-Turn, Multi-LLM Benchmark for LLM Agent Security","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CR","authors_text":"Chuan Li, David Hartmann, Devina Jain","submitted_at":"2026-07-20T15:30:38Z","abstract_excerpt":"LLM-based agents process external content, exposing them to prompt injection and multi-turn manipulation. Most safety benchmarks evaluate defenders against fixed attack pools collected before evaluation, single-turn or multi-turn. We present a 21-scenario benchmark for \\emph{adaptive multi-round attacks against memoryless LLM defenders}: an autonomous LLM attacker observes prior defender responses and pivots across rounds, while each defender response is evaluated as a fresh interaction. Holding the 21 scenarios, attackers, defenders, and structured-output scoring fixed, restricting scoring to"},"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":"2607.18063","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CR","submitted_at":"2026-07-20T15:30:38Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"a099a82d40bd41c0682983738eadebc12e80b81758653ba92681e6d3d8d1b420","abstract_canon_sha256":"7d7ababd08c217524789866343ed38ad5083e1ce4226012eda169c6c76aba2be"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-21T02:22:14.357490Z","signature_b64":"OlFfKpDmI8ZPhevDa03dMiyq2DHg++NbwSO7hjXIQuTAt7wf2In1OhOeyTMc8gR9Xm7PW91OASFr9rkLLOKPBA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"7b13ac22ab176e432636b98eb620cc4977e5a0990019a7cf5c80d10b580e2824","last_reissued_at":"2026-07-21T02:22:14.356645Z","signature_status":"signed_v1","first_computed_at":"2026-07-21T02:22:14.356645Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Adaptive Adversaries: A Multi-Turn, Multi-LLM Benchmark for LLM Agent Security","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CR","authors_text":"Chuan Li, David Hartmann, Devina Jain","submitted_at":"2026-07-20T15:30:38Z","abstract_excerpt":"LLM-based agents process external content, exposing them to prompt injection and multi-turn manipulation. Most safety benchmarks evaluate defenders against fixed attack pools collected before evaluation, single-turn or multi-turn. We present a 21-scenario benchmark for \\emph{adaptive multi-round attacks against memoryless LLM defenders}: an autonomous LLM attacker observes prior defender responses and pivots across rounds, while each defender response is evaluated as a fresh interaction. Holding the 21 scenarios, attackers, defenders, and structured-output scoring fixed, restricting scoring to"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2607.18063","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/2607.18063/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":"2607.18063","created_at":"2026-07-21T02:22:14.357081+00:00"},{"alias_kind":"arxiv_version","alias_value":"2607.18063v1","created_at":"2026-07-21T02:22:14.357081+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.18063","created_at":"2026-07-21T02:22:14.357081+00:00"},{"alias_kind":"pith_short_12","alias_value":"PMJ2YIVLC5XE","created_at":"2026-07-21T02:22:14.357081+00:00"},{"alias_kind":"pith_short_16","alias_value":"PMJ2YIVLC5XEGJRW","created_at":"2026-07-21T02:22:14.357081+00:00"},{"alias_kind":"pith_short_8","alias_value":"PMJ2YIVL","created_at":"2026-07-21T02:22:14.357081+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/PMJ2YIVLC5XEGJRWXGHLMIGMJF","json":"https://pith.science/pith/PMJ2YIVLC5XEGJRWXGHLMIGMJF.json","graph_json":"https://pith.science/api/pith-number/PMJ2YIVLC5XEGJRWXGHLMIGMJF/graph.json","events_json":"https://pith.science/api/pith-number/PMJ2YIVLC5XEGJRWXGHLMIGMJF/events.json","paper":"https://pith.science/paper/PMJ2YIVL"},"agent_actions":{"view_html":"https://pith.science/pith/PMJ2YIVLC5XEGJRWXGHLMIGMJF","download_json":"https://pith.science/pith/PMJ2YIVLC5XEGJRWXGHLMIGMJF.json","view_paper":"https://pith.science/paper/PMJ2YIVL","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2607.18063&json=true","fetch_graph":"https://pith.science/api/pith-number/PMJ2YIVLC5XEGJRWXGHLMIGMJF/graph.json","fetch_events":"https://pith.science/api/pith-number/PMJ2YIVLC5XEGJRWXGHLMIGMJF/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/PMJ2YIVLC5XEGJRWXGHLMIGMJF/action/timestamp_anchor","attest_storage":"https://pith.science/pith/PMJ2YIVLC5XEGJRWXGHLMIGMJF/action/storage_attestation","attest_author":"https://pith.science/pith/PMJ2YIVLC5XEGJRWXGHLMIGMJF/action/author_attestation","sign_citation":"https://pith.science/pith/PMJ2YIVLC5XEGJRWXGHLMIGMJF/action/citation_signature","submit_replication":"https://pith.science/pith/PMJ2YIVLC5XEGJRWXGHLMIGMJF/action/replication_record"}},"created_at":"2026-07-21T02:22:14.357081+00:00","updated_at":"2026-07-21T02:22:14.357081+00:00"}