MIV v0.2-freeze: Audit-ready evaluation runs for programme QA (provenance hardening)

## MIV-BRV Toolkit (v0.1) This record provides three citable research objects for auditing cross-LLM interpretation variance: 1) Interaction Log Schema v1.0 (JSONL/CSV header + field dictionary) 2) Run Card + Gate Specification v1.0 (6-field Run Card + G1–G5 Gates) 3) Minimal Benchmark Builder v1.0 (rules/templates to construct a 50-slice benchmark) ### ScopeEngineering-oriented auditing of LLM outputs via traceable evidence spans and contestability gates. ### Non-scopeNot a truth metric; not a causal law; not a claim of endorsement or integration with any third party. ### How to citeAfter Zenodo publication, cite the **version DOI** for reproducibility, or the **concept DOI** for the latest version: - Version DOI (v0.1): 10.5281/zenodo.18209567- Concept DOI (all versions): 10.5281/zenodo.18209566 - Hon Bor So. (2026). *MIV-BRV Toolkit: Run Cards, Gates, and Minimal Benchmark* (v0.1). Zenodo. 10.5281/zenodo.18209567 Author: Hon Bor So Note: The author also publishes under the name Sing So (Chöndrel Dorje). ORCID: 0009-0008-2768-7494 Related identifiers are cataloging metadata only and imply no endorsement.