{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:MLTYFDJFVCSODMBLKJCGDBB4UQ","short_pith_number":"pith:MLTYFDJF","schema_version":"1.0","canonical_sha256":"62e7828d25a8a4e1b02b524461843ca4219b865ecec75038ed9cc2bd1f9cfabc","source":{"kind":"arxiv","id":"2410.16155","version":2},"attestation_state":"computed","paper":{"title":"A Troublemaker with Contagious Jailbreak Makes Chaos in Honest Towns","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Jun Zhao, Kang Liu, Pengfei Cao, Tianyi Men, Yubo Chen, Zhuoran Jin","submitted_at":"2024-10-21T16:21:24Z","abstract_excerpt":"With the development of large language models, they are widely used as agents in various fields. A key component of agents is memory, which stores vital information but is susceptible to jailbreak attacks. Existing research mainly focuses on single-agent attacks and shared memory attacks. However, real-world scenarios often involve independent memory. In this paper, we propose the Troublemaker Makes Chaos in Honest Town (TMCHT) task, a large-scale, multi-agent, multi-topology text-based attack evaluation framework. TMCHT involves one attacker agent attempting to mislead an entire society of ag"},"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":"2410.16155","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-10-21T16:21:24Z","cross_cats_sorted":[],"title_canon_sha256":"43cba7c6c1ab689dd68153349c40742c9d3d9305782b86fdf84706b2f61d81df","abstract_canon_sha256":"09d288b61ee74cfb443729785fe5de08c37f5d41218a8dd8e63f077e6e6e6f66"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:27:20.994664Z","signature_b64":"xZmKczf2ZCzjPU9f0oHZ4zgtTl16Bok6AHgWwJ9GjTSvs0A17oLQtEdgq7tP/qCB+ktBw8UA/tT8Kzrfz8OHBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"62e7828d25a8a4e1b02b524461843ca4219b865ecec75038ed9cc2bd1f9cfabc","last_reissued_at":"2026-07-05T11:27:20.994190Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:27:20.994190Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"A Troublemaker with Contagious Jailbreak Makes Chaos in Honest Towns","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Jun Zhao, Kang Liu, Pengfei Cao, Tianyi Men, Yubo Chen, Zhuoran Jin","submitted_at":"2024-10-21T16:21:24Z","abstract_excerpt":"With the development of large language models, they are widely used as agents in various fields. A key component of agents is memory, which stores vital information but is susceptible to jailbreak attacks. Existing research mainly focuses on single-agent attacks and shared memory attacks. However, real-world scenarios often involve independent memory. In this paper, we propose the Troublemaker Makes Chaos in Honest Town (TMCHT) task, a large-scale, multi-agent, multi-topology text-based attack evaluation framework. TMCHT involves one attacker agent attempting to mislead an entire society of ag"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2410.16155","kind":"arxiv","version":2},"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/2410.16155/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":"2410.16155","created_at":"2026-07-05T11:27:20.994253+00:00"},{"alias_kind":"arxiv_version","alias_value":"2410.16155v2","created_at":"2026-07-05T11:27:20.994253+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2410.16155","created_at":"2026-07-05T11:27:20.994253+00:00"},{"alias_kind":"pith_short_12","alias_value":"MLTYFDJFVCSO","created_at":"2026-07-05T11:27:20.994253+00:00"},{"alias_kind":"pith_short_16","alias_value":"MLTYFDJFVCSODMBL","created_at":"2026-07-05T11:27:20.994253+00:00"},{"alias_kind":"pith_short_8","alias_value":"MLTYFDJF","created_at":"2026-07-05T11:27:20.994253+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2608.10218","citing_title":"Mind Viruses: Self-Propagating Ideas in Multi-Agent LLM Systems","ref_index":29,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/MLTYFDJFVCSODMBLKJCGDBB4UQ","json":"https://pith.science/pith/MLTYFDJFVCSODMBLKJCGDBB4UQ.json","graph_json":"https://pith.science/api/pith-number/MLTYFDJFVCSODMBLKJCGDBB4UQ/graph.json","events_json":"https://pith.science/api/pith-number/MLTYFDJFVCSODMBLKJCGDBB4UQ/events.json","paper":"https://pith.science/paper/MLTYFDJF"},"agent_actions":{"view_html":"https://pith.science/pith/MLTYFDJFVCSODMBLKJCGDBB4UQ","download_json":"https://pith.science/pith/MLTYFDJFVCSODMBLKJCGDBB4UQ.json","view_paper":"https://pith.science/paper/MLTYFDJF","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2410.16155&json=true","fetch_graph":"https://pith.science/api/pith-number/MLTYFDJFVCSODMBLKJCGDBB4UQ/graph.json","fetch_events":"https://pith.science/api/pith-number/MLTYFDJFVCSODMBLKJCGDBB4UQ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/MLTYFDJFVCSODMBLKJCGDBB4UQ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/MLTYFDJFVCSODMBLKJCGDBB4UQ/action/storage_attestation","attest_author":"https://pith.science/pith/MLTYFDJFVCSODMBLKJCGDBB4UQ/action/author_attestation","sign_citation":"https://pith.science/pith/MLTYFDJFVCSODMBLKJCGDBB4UQ/action/citation_signature","submit_replication":"https://pith.science/pith/MLTYFDJFVCSODMBLKJCGDBB4UQ/action/replication_record"}},"created_at":"2026-07-05T11:27:20.994253+00:00","updated_at":"2026-07-05T11:27:20.994253+00:00"}