SETT Framework: Technical Thesis and Architectural Position

SETT (Scalable Expert-based Task Topology) is a multi-agent AI framework designed around an explicit separation between probabilistic reasoning and operational authority. This technical report presents the current architectural thesis and engineering position of SETT at v0.12.0, including explicit role ownership, Actions as Data, deterministic governance, causal tracing, lifecycle control, cooperative cancellation, deadlines, structured terminal outcomes, and idempotency at the real-world effect boundary.

The report documents the selected architectural evolution of SETT, its current evidence base, known limitations, and the conditions under which its central thesis could be weakened or falsified. For this edition, the SETT v0.12.0 source archive was executed in a separate review environment and the complete test suite finished with 893/893 tests passing. This result demonstrates consistency with the tested contracts; it is not presented as a security certification, proof of universal correctness, or independent third-party validation.

SETT remains under active development. APIs, internal structures, mechanisms, and previous design decisions may be extended, revised, replaced, or reverted as testing, benchmarking, external review, and real downstream use provide new evidence. This report therefore represents a technical snapshot of SETT v0.12.0 rather than a declaration of final architecture.

The project's intended stable principles are that capability should not automatically imply authority; reasoning, authorization, and effect should remain distinguishable; governance should be explicit and auditable; and technical claims should remain open to verification and falsification. Meaningful implementation changes, completed work, known limitations, and planned evolution remain traceable through the public SETT repository and the 4HumanityAI project.

This record is the canonical English-language edition. Official Spanish and Japanese editions are published as separate language variants.

Related publication:
This technical report builds on and updates the architectural direction introduced in the earlier English-language SETT preprint: SETT: Scalable Expert-based Task Topology — A macro-scale application of the Badger Architecture philosophy to pre-designed multi-agent systems (DOI: 10.5281/zenodo.21301637). The present report focuses on the framework's current technical position, architectural evolution, governance boundaries, implementation evidence, and known limitations as of SETT v0.12.0.

Official language editions:
Spanish edition — SETT Framework: Tesis técnica y posición arquitectónica - DOI: 10.5281/zenodo.21927405
Japanese edition — SETT Framework: 技術的主張とアーキテクチャ方針 - DOI: 10.5281/zenodo.21927613