{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:JNUWB7NHBUWI3KR54HRL2WW6DG","short_pith_number":"pith:JNUWB7NH","schema_version":"1.0","canonical_sha256":"4b6960fda70d2c8daa3de1e2bd5ade1991cdb106d6a8c80e969005567a16c508","source":{"kind":"arxiv","id":"2508.18652","version":1},"attestation_state":"computed","paper":{"title":"UniC-RAG: Universal Knowledge Corruption Attacks to Retrieval-Augmented Generation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.CR","authors_text":"Jinyuan Jia, Runpeng Geng, Yanting Wang, Ying Chen","submitted_at":"2025-08-26T03:50:52Z","abstract_excerpt":"Retrieval-augmented generation (RAG) systems are widely deployed in real-world applications in diverse domains such as finance, healthcare, and cybersecurity. However, many studies showed that they are vulnerable to knowledge corruption attacks, where an attacker can inject adversarial texts into the knowledge database of a RAG system to induce the LLM to generate attacker-desired outputs. Existing studies mainly focus on attacking specific queries or queries with similar topics (or keywords). In this work, we propose UniC-RAG, a universal knowledge corruption attack against RAG systems. Unlik"},"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":"2508.18652","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CR","submitted_at":"2025-08-26T03:50:52Z","cross_cats_sorted":["cs.CL"],"title_canon_sha256":"b5ce7e953cd42ca1b245f1dd08fba2ec23aa9b006f3e62f28a0556ad9aaa777b","abstract_canon_sha256":"649df7421d6ca4aa548deb85616233de1797790fba931d2e44c5204318cb030e"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:59:28.636116Z","signature_b64":"ZJfaKkab1kQeWHu7e1ierFi06HDXt3bYeykDakAqpan5paLpcGIrFLgpMR1FJUfPZBww5r/oPsMuF0dV3yv6BA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"4b6960fda70d2c8daa3de1e2bd5ade1991cdb106d6a8c80e969005567a16c508","last_reissued_at":"2026-07-05T11:59:28.635652Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:59:28.635652Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"UniC-RAG: Universal Knowledge Corruption Attacks to Retrieval-Augmented Generation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.CR","authors_text":"Jinyuan Jia, Runpeng Geng, Yanting Wang, Ying Chen","submitted_at":"2025-08-26T03:50:52Z","abstract_excerpt":"Retrieval-augmented generation (RAG) systems are widely deployed in real-world applications in diverse domains such as finance, healthcare, and cybersecurity. However, many studies showed that they are vulnerable to knowledge corruption attacks, where an attacker can inject adversarial texts into the knowledge database of a RAG system to induce the LLM to generate attacker-desired outputs. Existing studies mainly focus on attacking specific queries or queries with similar topics (or keywords). In this work, we propose UniC-RAG, a universal knowledge corruption attack against RAG systems. Unlik"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2508.18652","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/2508.18652/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":"2508.18652","created_at":"2026-07-05T11:59:28.635716+00:00"},{"alias_kind":"arxiv_version","alias_value":"2508.18652v1","created_at":"2026-07-05T11:59:28.635716+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2508.18652","created_at":"2026-07-05T11:59:28.635716+00:00"},{"alias_kind":"pith_short_12","alias_value":"JNUWB7NHBUWI","created_at":"2026-07-05T11:59:28.635716+00:00"},{"alias_kind":"pith_short_16","alias_value":"JNUWB7NHBUWI3KR5","created_at":"2026-07-05T11:59:28.635716+00:00"},{"alias_kind":"pith_short_8","alias_value":"JNUWB7NH","created_at":"2026-07-05T11:59:28.635716+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":5,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.21377","citing_title":"ARENA: An Architecture for Measuring the Transferability of Autonomous Cyber Defense","ref_index":37,"is_internal_anchor":false},{"citing_arxiv_id":"2606.03354","citing_title":"ImageAuditor: Membership Inference Attack against Image-based Retrieval-Augmented Generation","ref_index":10,"is_internal_anchor":false},{"citing_arxiv_id":"2606.01212","citing_title":"DiscourseFlip: An Oblique Discourse-Level Opinion Manipulation Attack against Black-box Retrieval-Augmented Generation","ref_index":10,"is_internal_anchor":false},{"citing_arxiv_id":"2605.28112","citing_title":"A Wolf in Sheep's Clothing: Targeted Routing Hijacking in Federated RAG","ref_index":4,"is_internal_anchor":false},{"citing_arxiv_id":"2604.05480","citing_title":"Can You Trust the Vectors in Your Vector Database? Black-Hole Attack from Embedding Space Defects","ref_index":11,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/JNUWB7NHBUWI3KR54HRL2WW6DG","json":"https://pith.science/pith/JNUWB7NHBUWI3KR54HRL2WW6DG.json","graph_json":"https://pith.science/api/pith-number/JNUWB7NHBUWI3KR54HRL2WW6DG/graph.json","events_json":"https://pith.science/api/pith-number/JNUWB7NHBUWI3KR54HRL2WW6DG/events.json","paper":"https://pith.science/paper/JNUWB7NH"},"agent_actions":{"view_html":"https://pith.science/pith/JNUWB7NHBUWI3KR54HRL2WW6DG","download_json":"https://pith.science/pith/JNUWB7NHBUWI3KR54HRL2WW6DG.json","view_paper":"https://pith.science/paper/JNUWB7NH","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2508.18652&json=true","fetch_graph":"https://pith.science/api/pith-number/JNUWB7NHBUWI3KR54HRL2WW6DG/graph.json","fetch_events":"https://pith.science/api/pith-number/JNUWB7NHBUWI3KR54HRL2WW6DG/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/JNUWB7NHBUWI3KR54HRL2WW6DG/action/timestamp_anchor","attest_storage":"https://pith.science/pith/JNUWB7NHBUWI3KR54HRL2WW6DG/action/storage_attestation","attest_author":"https://pith.science/pith/JNUWB7NHBUWI3KR54HRL2WW6DG/action/author_attestation","sign_citation":"https://pith.science/pith/JNUWB7NHBUWI3KR54HRL2WW6DG/action/citation_signature","submit_replication":"https://pith.science/pith/JNUWB7NHBUWI3KR54HRL2WW6DG/action/replication_record"}},"created_at":"2026-07-05T11:59:28.635716+00:00","updated_at":"2026-07-05T11:59:28.635716+00:00"}