{
  "id": 4022399,
  "title": "rag-sanitizer v0.2: RAG pre-ingestion corpus scanner with real embeddings and CLIP",
  "url": "https://urgent.news/2026/08/28/rag-sanitizer-v0-2-escaner-de-corpus-rag-pre-ingestion-con-embeddings",
  "topic": "ai",
  "section": "AI",
  "published": "2026-08-28T19:21:34.000Z",
  "source": {
    "name": "Dev.to",
    "slug": "dev-to",
    "url": "https://dev.to/magopredator/rag-sanitizer-v02-escaner-de-corpus-rag-pre-ingestion-con-embeddings-reales-y-clip-3jo4"
  },
  "original_language": "es",
  "account": "The rag-sanitizer v0.2 is a scanner for detecting poisoned documents in a Retrieval-Augmented Generation (RAG) corpus before ingestion. It uses real embeddings and CLIP to identify documents that mimic the style of the corpus, swap entities, or contain images that contradict the text. The scanner outputs a report for each document, labeling it as CLEAN, SUSPECT, or POISON, along with the reasons. The tool is available on GitHub under the AGPL-3.0-or-later license.",
  "summary": "rag-sanitizer v0.2: escáner de corpus RAG pre-ingestión con embeddings reales y CLIP Detecta documentos poison en un corpus RAG antes del chunk+embed, no después de la recuperación. v0.2 añade embeddings reales (MiniLM) y detección multimodal (CLIP). El problema La mayoría de la defensa contra envenenamiento de RAG opera post-retrieval o en inferencia . Pero el corpus se puede envenar antes de la…",
  "key_points": [],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}