Method Definition v1.0

Method Definition v1.0 formalizes a method for structuring practical AI instruction in which a general-purpose conversational LLM serves as the primary learning environment. The method combines six mandatory functions: an author-governed instructional specification; an instruction-programmed AI teacher; runtime micro-adaptation; learner steering; evidence-based checking and remediation; and transfer to the learner's own real tasks. The document also defines two assembly principles — Balancing and Layered Failure-Mode Coverage — as well as the method's Minimal Profile, the boundary between the method and its implementation or content, and the claims that are explicitly not made. This record contains the canonical English version and an authorized Russian translation in PDF and Markdown formats. It also includes SHA256SUMS.txt containing SHA-256 hashes of the four content files. Version 1.0 was finalized and first published on 14 August 2026. Supporting evidence, development chronology, prior-art evaluation, pilot results, external research, and review records are maintained separately and are not reproduced in this deposit.