Metraipolis › archive › dc_3efc643ebe2b

Control-Data Flow Separation: Stable Prompt Optimization in Multi-Agent LLMs — arXiv:2609.00621

Source: arXiv (en)
Original: https://arxiv.org/abs/2609.00621
License (as archived): CC0-1.0
Archived: 2026-10-17T06:30:48.774Z
SHA-256 of the archived text: bb6343a4bbed64d6819f84779b52a504467696d21ce728ae869ce4557a29f9d2
Stable URL of this page: https://metraipolis.com/documento/dc_3efc643ebe2b

How to cite

Control-Data Flow Separation: Stable Prompt Optimization in Multi-Agent LLMs — arXiv:2609.00621 — arXiv (en), https://arxiv.org/abs/2609.00621 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_3efc643ebe2b (Metraipolis archive dc_3efc643ebe2b, sha256 bb6343a4bbed64d6819f84779b52a504467696d21ce728ae869ce4557a29f9d2)

BibTeX

@misc{dc3efc643ebe2b,
  title = {Control-Data Flow Separation: Stable Prompt Optimization in Multi-Agent LLMs — arXiv:2609.00621},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.00621},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_3efc643ebe2b; sha256 bb6343a4bbed64d6819f84779b52a504467696d21ce728ae869ce4557a29f9d2},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.00621. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Prompt optimization can improve multi-agent LLM systems, but the prompts being optimized often serve two entangled roles: generating task-relevant content and specifying execution-critical protocols, such as message routing, output formatting, and termination signals, on which the underlying code relies. As a result, a prompt edit intended to improve content generation can inadvertently corrupt the protocol and cause the entire agent pipeline to fail. Our key observation is that these two roles have different representations: execution protocols are typically structured, while task-relevant content is usually expressed in unstructured language. Based on this, we propose control-data flow separation, where execution-critical control is represented as typed, validated program objects, while task-relevant language remains the optimizable data flow for agent communication. This design allows optimizers to improve multi-agent behavior without exposing the routing or formatting interface to prompt drift. Across synthetic reasoning, collaborative review generation, and insurance rating workflows, our framework empirically achieves 100% eventual protocol validity while consistently improving task performance.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-02 in cs.MA).
Categorie arXiv: Artificial Intelligence (cs.AI) ; Computation and Language (cs.CL); Multiagent Systems (cs.MA).
Pagina del lavoro: https://arxiv.org/abs/2609.00621 — 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 2026-10-17T06:14:52.157Z dalla pagina https://arxiv.org/catchup/cs.MA/2026-09-02?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).