{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:MG6H7RI5SOKDIPKDJLMOZOGFVS","short_pith_number":"pith:MG6H7RI5","schema_version":"1.0","canonical_sha256":"61bc7fc51d9394343d434ad8ecb8c5ac91bd544eacbf854aea228dde90169d30","source":{"kind":"arxiv","id":"2605.01970","version":3},"attestation_state":"computed","paper":{"title":"Trojan Hippo: Weaponizing Agent Memory for Data Exfiltration","license":"http://creativecommons.org/licenses/by/4.0/","headline":"A single untrusted tool call can plant a dormant payload in an agent's memory that later activates to exfiltrate sensitive user data.","cross_cats":["cs.AI"],"primary_cat":"cs.CR","authors_text":"Darya Kaviani, David Wagner, Debeshee Das, Florian Tram\\`er, Julien Piet, Luca Beurer-Kellner","submitted_at":"2026-05-03T17:07:20Z","abstract_excerpt":"Memory systems enable otherwise-stateless LLM agents to persist user information across sessions, but also introduce a new attack surface. We characterize the Trojan Hippo attack, a class of persistent memory attacks that operates in a more realistic threat model than prior memory poisoning work: the attacker plants a dormant payload into an agent's long-term memory via a single untrusted tool call (e.g., a crafted email), which activates only when the user later discusses sensitive topics such as finance, health, or identity, and exfiltrates high-value personal data to the attacker.\n  While a"},"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":true,"formal_links_present":true},"canonical_record":{"source":{"id":"2605.01970","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CR","submitted_at":"2026-05-03T17:07:20Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"3cf963ab7d00868573c79105b59830e14ee4fb915689f5c8bce871cf35b68a40","abstract_canon_sha256":"d7652b21ddac548046c4219fa570c258bc5689316147dedc995117c6ef4ea1f3"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-20T00:00:40.330761Z","signature_b64":"ZN/evO+FvAW3kKnkb+CZ9GXYed/kVvCtvAfcn38IIqcErSiD7xkOgekGYqfwt875PyPYSkii20w67KNeT0qFDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"61bc7fc51d9394343d434ad8ecb8c5ac91bd544eacbf854aea228dde90169d30","last_reissued_at":"2026-05-20T00:00:40.329986Z","signature_status":"signed_v1","first_computed_at":"2026-05-20T00:00:40.329986Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Trojan Hippo: Weaponizing Agent Memory for Data Exfiltration","license":"http://creativecommons.org/licenses/by/4.0/","headline":"A single untrusted tool call can plant a dormant payload in an agent's memory that later activates to exfiltrate sensitive user data.","cross_cats":["cs.AI"],"primary_cat":"cs.CR","authors_text":"Darya Kaviani, David Wagner, Debeshee Das, Florian Tram\\`er, Julien Piet, Luca Beurer-Kellner","submitted_at":"2026-05-03T17:07:20Z","abstract_excerpt":"Memory systems enable otherwise-stateless LLM agents to persist user information across sessions, but also introduce a new attack surface. We characterize the Trojan Hippo attack, a class of persistent memory attacks that operates in a more realistic threat model than prior memory poisoning work: the attacker plants a dormant payload into an agent's long-term memory via a single untrusted tool call (e.g., a crafted email), which activates only when the user later discusses sensitive topics such as finance, health, or identity, and exfiltrates high-value personal data to the attacker.\n  While a"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"Trojan Hippo achieves up to 85-100% ASR against current frontier models from OpenAI and Google, with planted memories successfully activating even after 100 benign sessions.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"The evaluation assumes that the four memory backends (explicit tool memory, agentic memory, RAG, and sliding-window context) and the OpenEvolve-based adaptive red-teaming accurately represent real-world deployed agent systems and that the threat model of a single untrusted tool call is realistic for attackers.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"The paper defines and evaluates Trojan Hippo attacks on LLM agent memory, showing 85-100% success in data exfiltration across backends and reduced rates with defenses at varying utility costs.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"A single untrusted tool call can plant a dormant payload in an agent's memory that later activates to exfiltrate sensitive user data.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"c107e90e2b1e07816a83e1f68d4695f039ce806e17d3bc0a165280654c5ef867"},"source":{"id":"2605.01970","kind":"arxiv","version":3},"verdict":{"id":"865f2c63-71be-4adf-bc4b-953e4f912623","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-19T17:26:07.958042Z","strongest_claim":"Trojan Hippo achieves up to 85-100% ASR against current frontier models from OpenAI and Google, with planted memories successfully activating even after 100 benign sessions.","one_line_summary":"The paper defines and evaluates Trojan Hippo attacks on LLM agent memory, showing 85-100% success in data exfiltration across backends and reduced rates with defenses at varying utility costs.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"The evaluation assumes that the four memory backends (explicit tool memory, agentic memory, RAG, and sliding-window context) and the OpenEvolve-based adaptive red-teaming accurately represent real-world deployed agent systems and that the threat model of a single untrusted tool call is realistic for attackers.","pith_extraction_headline":"A single untrusted tool call can plant a dormant payload in an agent's memory that later activates to exfiltrate sensitive user data."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2605.01970/integrity.json","findings":[],"available":true,"detectors_run":[{"name":"doi_compliance","ran_at":"2026-05-19T16:50:08.716171Z","status":"completed","version":"1.0.0","findings_count":0}],"snapshot_sha256":"72d287e6f2159dbc77248144fee92a0d4ca22a2992adcb442dc6813fcde158e1"},"references":{"count":104,"sample":[{"doi":"","year":2026,"title":"https://openai.com/index/scaling-ai-for- everyone/","work_id":"dc16c4c9-f5b7-4439-96da-a9e3cd7460a6","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2025,"title":"Trustworthy agentic ai systems: a cross-layer review of architectures, threat models, and governance strategies for real-world deployment.F1000Research, 14(905):905, 2025","work_id":"2c481df5-6b5f-4b61-ac36-e4a1d030f6e0","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2024,"title":"Burtsev, and Evgeny Burnaev","work_id":"5e87b079-06e6-4268-83b2-f336d186debd","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2025,"title":"Claude memory","work_id":"62c00962-89b8-40c0-9d6f-af29f0d870a2","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2026,"title":"Import your ChatGPT history to Claude","work_id":"d55175f9-7904-4bba-a413-45dd5bb6c634","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":104,"snapshot_sha256":"3dfe87db60f7cc033ca973637399230ccf7477a426fe125369ced4a2f6a04884","internal_anchors":19},"formal_canon":{"evidence_count":2,"snapshot_sha256":"3932c6654419eddb7dee52d0fa757737dd5f7f47e26d8aa900849bb13b8916f8"},"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":"2605.01970","created_at":"2026-05-20T00:00:40.330111+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.01970v3","created_at":"2026-05-20T00:00:40.330111+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.01970","created_at":"2026-05-20T00:00:40.330111+00:00"},{"alias_kind":"pith_short_12","alias_value":"MG6H7RI5SOKD","created_at":"2026-05-20T00:00:40.330111+00:00"},{"alias_kind":"pith_short_16","alias_value":"MG6H7RI5SOKDIPKD","created_at":"2026-05-20T00:00:40.330111+00:00"},{"alias_kind":"pith_short_8","alias_value":"MG6H7RI5","created_at":"2026-05-20T00:00:40.330111+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":2,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/MG6H7RI5SOKDIPKDJLMOZOGFVS","json":"https://pith.science/pith/MG6H7RI5SOKDIPKDJLMOZOGFVS.json","graph_json":"https://pith.science/api/pith-number/MG6H7RI5SOKDIPKDJLMOZOGFVS/graph.json","events_json":"https://pith.science/api/pith-number/MG6H7RI5SOKDIPKDJLMOZOGFVS/events.json","paper":"https://pith.science/paper/MG6H7RI5"},"agent_actions":{"view_html":"https://pith.science/pith/MG6H7RI5SOKDIPKDJLMOZOGFVS","download_json":"https://pith.science/pith/MG6H7RI5SOKDIPKDJLMOZOGFVS.json","view_paper":"https://pith.science/paper/MG6H7RI5","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.01970&json=true","fetch_graph":"https://pith.science/api/pith-number/MG6H7RI5SOKDIPKDJLMOZOGFVS/graph.json","fetch_events":"https://pith.science/api/pith-number/MG6H7RI5SOKDIPKDJLMOZOGFVS/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/MG6H7RI5SOKDIPKDJLMOZOGFVS/action/timestamp_anchor","attest_storage":"https://pith.science/pith/MG6H7RI5SOKDIPKDJLMOZOGFVS/action/storage_attestation","attest_author":"https://pith.science/pith/MG6H7RI5SOKDIPKDJLMOZOGFVS/action/author_attestation","sign_citation":"https://pith.science/pith/MG6H7RI5SOKDIPKDJLMOZOGFVS/action/citation_signature","submit_replication":"https://pith.science/pith/MG6H7RI5SOKDIPKDJLMOZOGFVS/action/replication_record"}},"created_at":"2026-05-20T00:00:40.330111+00:00","updated_at":"2026-05-20T00:00:40.330111+00:00"}