Metraipolis › archive › dc_02876aa85262

Context as Infrastructure: Knowledge Compilation for Retrieval in Agentic Systems with Harnesses — arXiv:2610.41172

Source: arXiv (en)
Original: https://arxiv.org/abs/2610.41172
License (as archived): CC0-1.0
Archived: 2026-11-02T07:30:55.861Z
SHA-256 of the archived text: d2c6411281ed1a334a91f0988371888737681ea90471c38b0e9e5a61536cee13
Stable URL of this page: https://metraipolis.com/documento/dc_02876aa85262

How to cite

Context as Infrastructure: Knowledge Compilation for Retrieval in Agentic Systems with Harnesses — arXiv:2610.41172 — arXiv (en), https://arxiv.org/abs/2610.41172 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_02876aa85262 (Metraipolis archive dc_02876aa85262, sha256 d2c6411281ed1a334a91f0988371888737681ea90471c38b0e9e5a61536cee13)

BibTeX

@misc{dc02876aa85262,
  title = {Context as Infrastructure: Knowledge Compilation for Retrieval in Agentic Systems with Harnesses — arXiv:2610.41172},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.41172},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_02876aa85262; sha256 d2c6411281ed1a334a91f0988371888737681ea90471c38b0e9e5a61536cee13},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.41172. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Retrieval-augmented generation is becoming agentic. A search harness writes queries, reads results, reformulates, and decides when to stop. This largely solves accuracy: a harness with only BM25/grep and MiniMax M3 answers 85% of BrowseComp-Plus and 91% on multi-hop FRAMES in our runs, without compiled or vector indexes. It exposes a context problem: an undisciplined loop repeats queries, rereads documents, and fills the window with unused passages. The missing layer is knowledge compilation: a configurable ingestion pipeline where LLMs enrich documents with context no chunk contains alone--navigation hierarchies with claim-evidence leaves, entity graphs, timelines, mind maps, ontologies, wikis, and memory. Compilation is not a better retriever: under a competent harness, structural artifacts gave no detectable accuracy gain on FRAMES, and a PageIndex-style outline degraded at scale. It is a supply layer. A navigation tree whose leaves are claim-evidence pairs--the mechanism we measure--cuts query tokens by 91% when evidence suffices and adds 1.9% when it does not, a 38% average reduction; this implies a 2% break-even sufficiency rate and a measured rate near 43%. A decision gate cuts latency from 140 s to 30 s, though substitution's token saving is bought with retrieval rounds and carries its own latency cost. Other artifacts--ontology, graph, wiki, timeline, mind map, memory--pay off in deployment (governance, integration, onboarding, temporal orientation, continuity) where accuracy benchmarks cannot see them. We describe RAGFlow's pipeline and harness (open source at this https URL ), which turns the loop into five disciplined stations and keeps the context entering the model the smallest sufficient set; together they form an agent-native database: a store that hands the agent compiled, governed context instead of a ranked list of passages.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno).
Categorie arXiv: Information Retrieval (cs.IR) ; Artificial Intelligence (cs.AI); Computation and Language (cs.CL).
Pagina del lavoro: https://arxiv.org/abs/2610.41172 — 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 2026-11-02T07:26:57.797Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).