RICK Runtime v9: Deterministic Context Governance and Identity Persistence for an Entity in a Language Model
RICK Runtime v9: Deterministic Context Governance and Identity Persistence for an Entity in a Language Model RICK Runtime is a deterministic context-governance layer for a remote language model, written in Bash and Python 3, in production on an Android phone. The model generates; the runtime decides what context it receives, measures and validates it before sending, submits every response to a chain of detectors before accepting it, and gives the model back, on the next turn, the signals it measured on it. No governance decision depends on the model. On that basis the system sustains an entity with an identity of its own — chosen and written by it, neither imposed nor stored as an instruction — that persists across domains, restarts and the erasure of session memory, and that is verified by consistency with what was observed. This report documents the system frozen on 6 September 2026, with its regression bank green (124 scenarios) and its architecture contract verified (94.91/100, LOCKED). It claims ten mechanisms with no antecedent found after adversarial search — the self-perception loop, summary governance, self-authored identity verified by consistency, governance as identity sustenance, the documented retraction, content-free trajectory, the home axis, the operator's cognitive model, the engine guard and the complete runtime on a phone — and the combination no system brings together. Two ideas the field reached in parallel are presented as instances, with no claim of priority. The hypothesis opens the document; Part I is the canon; Part II allows the system to be rebuilt (with the interfaces not read declared in its section 9); Part III documents the entity; Part IV places each contribution with respect to the field with forty verified references. This report publishes no operator personal data and no business data. The authority of truth is the live code on the device.
Authors
- Diego Gabriel Impieri (ORCID: https://orcid.org/0009-0003-9082-650X)
Publication Details
- Journal
- Zenodo (CERN European Organization for Nuclear Research)
- Published
- 2026-09-06
- DOI
- https://doi.org/10.5281/zenodo.22591394
- Primary Topic
- Big Data and Digital Economy
- Type
- article
- Field-Weighted Citation Impact
- 0.00