Metraipolis › archive › dc_735ff630afc2

Reconcile Once, Write Anytime: A Trust-Tiered Librarian and a Multi-Agent Writer for Drift-Free, Point-in-Time Research — arXiv:2608.12984

Source
arXiv (en)
Original
https://arxiv.org/abs/2608.12984
License (as archived)
CC0-1.0
Archived
2026-10-19T04:16:39.954Z
SHA-256 of the archived text
395a892186425f53ba17a47b67a97f7c665c04fc50cee80feda26a7203ccd713
Stable URL of this page
https://metraipolis.com/documento/dc_735ff630afc2

How to cite

Reconcile Once, Write Anytime: A Trust-Tiered Librarian and a Multi-Agent Writer for Drift-Free, Point-in-Time Research — arXiv:2608.12984 — arXiv (en), https://arxiv.org/abs/2608.12984 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_735ff630afc2 (Metraipolis archive dc_735ff630afc2, sha256 395a892186425f53ba17a47b67a97f7c665c04fc50cee80feda26a7203ccd713)
BibTeX
@misc{dc735ff630afc2,
  title = {Reconcile Once, Write Anytime: A Trust-Tiered Librarian and a Multi-Agent Writer for Drift-Free, Point-in-Time Research — arXiv:2608.12984},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2608.12984},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_735ff630afc2; sha256 395a892186425f53ba17a47b67a97f7c665c04fc50cee80feda26a7203ccd713},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.12984. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Long-form research reports generated by large language models drift, contradict themselves, and lose provenance: the same metric appears with different values, and rumor is quoted as confidently as an audited filing. We present a two-tier agentic system that separates a maintained, point-in-time knowledge library from report writing. A deterministic "librarian" ingests timestamped sources into a trust-tiered ontology, layering evidence cards, an authoritative metric ledger, and a claim graph into an always-current source of truth, not per-query RAG over raw chunks. A portable multi-agent "writer" runtime then composes a contradiction-free, evidence-grounded report at any knowledge cutoff T, reading only evidence with as_of <= T (no look-ahead); red-team verdicts flow back into the librarian. We evaluate on a self-collected, public corpus of 6,130 sources yielding 555,926 evidence cards (SEC EDGAR filings across 295 issuers and 11 sectors, U.S. Bureau of Labor Statistics releases, and Wikipedia). From the one library we compose four point-in-time reports on distinct theses and run eight reproducible experiments, whose headline metrics come from a deterministic quality-control gate, itself validated by defect-injection meta-evaluation at recall 1.0 and precision 1.0. A shared metric ledger removes 6,845 cross-section contradictions to zero. Tier-first selection is correct on 22/22 gold cases where a popularity-first baseline scores only 9/22; trust tiering leaks zero media-sourced numbers, and no government statistic displaces a company's own filing. A red-team refutation propagates back and self-corrects a later run with zero manual edits. Replay exhibits zero look-ahead violations across seven cutoffs while the library grows from 235,373 to 555,312 cards. Difficulty-tiered model routing exceeds the all-Opus quality ceiling while running 3.7x faster than serial.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-14 in cs.MA).
Categorie arXiv: Multiagent Systems (cs.MA) ; Computation and Language (cs.CL).
Pagina del lavoro: https://arxiv.org/abs/2608.12984 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-19T04:12:45.945Z dalla pagina https://arxiv.org/catchup/cs.MA/2026-08-14?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).