Conflict-Aware Retriever Editing for Knowledge Injection Attacks on LLM-Based RAG Systems

arXiv 2606.18310•91465eaccf21be5bf46ccd1e0621b79408f4459f11695873c21419b8974c62f3
AgentraCAREATTACKGDPRHIPAAHNSWPII-exposureRAGSafeClawBenchTIGERVec2Textauditabilitydata-recoverydifferential-privacyembedding-inversionepoch-key-rotationfederated-learninggradient-inversionintrusion-response','IRSmemory-poisoningmodel-centric-attackprompt-injectionretriever-editingsoft-deletetool-using-agentsvector-database

Paper metadata

arXiv ID
2606.18310
Version
Not specified by this published record
Category
Computer Science — Cryptography and Security (cs.CR)

The PDF link points to arxiv.org. Baitaphish does not expose a private stored PDF.

Evidence and limitations

Source ID
arxiv_cs_cr
Record identifier
91465eaccf21be5bf46ccd1e0621b79408f4459f11695873c21419b8974c62f3
Enrichment time
2026-06-18T07:23:30Z
AI-assisted enrichment
Yes

This record may overlap with other records. Its enrichment can be incomplete or wrong, and machine assistance was used. Validate consequential decisions against the linked source and your own environment.