{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:TZ6ZCWUWEYJ2XMXW2K5QDWVXWC","short_pith_number":"pith:TZ6ZCWUW","schema_version":"1.0","canonical_sha256":"9e7d915a962613abb2f6d2bb01dab7b09b404d0ffd09596df74aa901aba5e37c","source":{"kind":"arxiv","id":"2408.04870","version":5},"attestation_state":"computed","paper":{"title":"ConfusedPilot: Confused Deputy Risks in RAG-based LLMs","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CR","authors_text":"Ayush RoyChowdhury, Mohit Tiwari, Mulong Luo, Prateek Sahu, Sarbartha Banerjee","submitted_at":"2024-08-09T05:20:05Z","abstract_excerpt":"Retrieval augmented generation (RAG) is a process where a large language model (LLM) retrieves useful information from a database and then generates the responses. It is becoming popular in enterprise settings for daily business operations. For example, Copilot for Microsoft 365 has accumulated millions of businesses. However, the security implications of adopting such RAG-based systems are unclear.\n  In this paper, we introduce ConfusedPilot, a class of security vulnerabilities of RAG systems that confuse Copilot and cause integrity and confidentiality violations in its responses. First, we i"},"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":"2408.04870","kind":"arxiv","version":5},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CR","submitted_at":"2024-08-09T05:20:05Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"62969920f416a4e4dc77feda8d7c5e2ed9337870eb9a2ae9b4a76bfd61035112","abstract_canon_sha256":"7297ee4a3865c028559dd5ede0ae211576e2b810ec6c28578f9b708734202c08"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:24:17.699860Z","signature_b64":"Ypvofr9fZfhSOFaAU6ZFjsGoNNkUkVjSdyqG9G18N+I4WpUAtIunP10qJh9+h1qluW+g8uqLiYsqJ+uNNsPUBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"9e7d915a962613abb2f6d2bb01dab7b09b404d0ffd09596df74aa901aba5e37c","last_reissued_at":"2026-07-05T09:24:17.699380Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:24:17.699380Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"ConfusedPilot: Confused Deputy Risks in RAG-based LLMs","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CR","authors_text":"Ayush RoyChowdhury, Mohit Tiwari, Mulong Luo, Prateek Sahu, Sarbartha Banerjee","submitted_at":"2024-08-09T05:20:05Z","abstract_excerpt":"Retrieval augmented generation (RAG) is a process where a large language model (LLM) retrieves useful information from a database and then generates the responses. It is becoming popular in enterprise settings for daily business operations. For example, Copilot for Microsoft 365 has accumulated millions of businesses. However, the security implications of adopting such RAG-based systems are unclear.\n  In this paper, we introduce ConfusedPilot, a class of security vulnerabilities of RAG systems that confuse Copilot and cause integrity and confidentiality violations in its responses. First, we i"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2408.04870","kind":"arxiv","version":5},"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/2408.04870/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":"2408.04870","created_at":"2026-07-05T09:24:17.699438+00:00"},{"alias_kind":"arxiv_version","alias_value":"2408.04870v5","created_at":"2026-07-05T09:24:17.699438+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2408.04870","created_at":"2026-07-05T09:24:17.699438+00:00"},{"alias_kind":"pith_short_12","alias_value":"TZ6ZCWUWEYJ2","created_at":"2026-07-05T09:24:17.699438+00:00"},{"alias_kind":"pith_short_16","alias_value":"TZ6ZCWUWEYJ2XMXW","created_at":"2026-07-05T09:24:17.699438+00:00"},{"alias_kind":"pith_short_8","alias_value":"TZ6ZCWUW","created_at":"2026-07-05T09:24:17.699438+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":8,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.18310","citing_title":"Conflict-Aware Retriever Editing for Knowledge Injection Attacks on LLM-Based RAG Systems","ref_index":36,"is_internal_anchor":false},{"citing_arxiv_id":"2605.12015","citing_title":"SkillSafetyBench: Evaluating Agent Safety under Skill-Facing Attack Surfaces","ref_index":30,"is_internal_anchor":false},{"citing_arxiv_id":"2605.14859","citing_title":"Do Coding Agents Understand Least-Privilege Authorization?","ref_index":7,"is_internal_anchor":false},{"citing_arxiv_id":"2603.12230","citing_title":"Security Considerations for Artificial Intelligence Agents","ref_index":46,"is_internal_anchor":false},{"citing_arxiv_id":"2604.00387","citing_title":"RAGShield: Detecting Numerical Claim Manipulation in Government RAG Systems","ref_index":12,"is_internal_anchor":false},{"citing_arxiv_id":"2605.12015","citing_title":"SkillSafetyBench: Evaluating Agent Safety under Skill-Facing Attack Surfaces","ref_index":75,"is_internal_anchor":false},{"citing_arxiv_id":"2605.09822","citing_title":"Oracle Poisoning: Corrupting Knowledge Graphs to Weaponise AI Agent Reasoning","ref_index":21,"is_internal_anchor":false},{"citing_arxiv_id":"2605.09594","citing_title":"Trust Me, Import This: Dependency Steering Attacks via Malicious Agent Skills","ref_index":29,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/TZ6ZCWUWEYJ2XMXW2K5QDWVXWC","json":"https://pith.science/pith/TZ6ZCWUWEYJ2XMXW2K5QDWVXWC.json","graph_json":"https://pith.science/api/pith-number/TZ6ZCWUWEYJ2XMXW2K5QDWVXWC/graph.json","events_json":"https://pith.science/api/pith-number/TZ6ZCWUWEYJ2XMXW2K5QDWVXWC/events.json","paper":"https://pith.science/paper/TZ6ZCWUW"},"agent_actions":{"view_html":"https://pith.science/pith/TZ6ZCWUWEYJ2XMXW2K5QDWVXWC","download_json":"https://pith.science/pith/TZ6ZCWUWEYJ2XMXW2K5QDWVXWC.json","view_paper":"https://pith.science/paper/TZ6ZCWUW","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2408.04870&json=true","fetch_graph":"https://pith.science/api/pith-number/TZ6ZCWUWEYJ2XMXW2K5QDWVXWC/graph.json","fetch_events":"https://pith.science/api/pith-number/TZ6ZCWUWEYJ2XMXW2K5QDWVXWC/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/TZ6ZCWUWEYJ2XMXW2K5QDWVXWC/action/timestamp_anchor","attest_storage":"https://pith.science/pith/TZ6ZCWUWEYJ2XMXW2K5QDWVXWC/action/storage_attestation","attest_author":"https://pith.science/pith/TZ6ZCWUWEYJ2XMXW2K5QDWVXWC/action/author_attestation","sign_citation":"https://pith.science/pith/TZ6ZCWUWEYJ2XMXW2K5QDWVXWC/action/citation_signature","submit_replication":"https://pith.science/pith/TZ6ZCWUWEYJ2XMXW2K5QDWVXWC/action/replication_record"}},"created_at":"2026-07-05T09:24:17.699438+00:00","updated_at":"2026-07-05T09:24:17.699438+00:00"}