User-Side Contextual Interaction Assessment (USCI) – Methodology Specification v1.0.0

User-Side Contextual Interaction Assessment (USCI) is a pre-empirical methodology specification for detecting and assessing User-Side Contextual Hallucination (USCH) in human–AI conversational interaction. USCI provides a four-axis risk assessment framework (FR/CA/SR/SA, scaled 0–4) with structured output schema, evidence-traceability rules, collapse-flag logic, and a post-assessment intervention recommendation layer (VCD). The specification separates a publicly auditable structural layer from controlled scoring internals to prevent adversarial gaming while enabling independent scrutiny. This deposit contains: USCI_Public_Methodology_v1.0.0_EN.pdf — Main methodology specification (19 pages) USCI_Methodology_v1.0.0_Official_Corrigendum_and_Compliance_Addendum_EN.pdf — Official Corrigendum and Compliance Addendum (3 pages) USCI_v1.0.0_PDF_ONLY_ZENODO_UPLOAD_READY.zip — Complete package with all files and SHA256 checksums SHA256_CHECKSUMS.txt — Integrity verification checksums This deposit is a public-layer, pre-empirical methodology release. Reproducibility in this record refers to structural auditability (definitions, output schema logic, evidence-traceability rules, and responsible-use boundaries), not deterministic reconstruction of controlled internals. This work is not a clinical or diagnostic instrument and must not be used as the sole basis for legal, punitive, or high-stakes determinations.