This replication package accompanies the article “Perceived skill mismatch as relative self-assessment: Personality, assessed competencies and job conditions”. It contains Stata/SE 19.5 do-files and Mata routines, non-disclosive aggregate benchmarks, table-data files, Figure 1 code and outputs, and validation documentation. The code reconstructs the analysis from OECD Survey of Adult Skills 2023 (PIAAC Cycle 2) country public-use files, forms deterministic estimation, development and held-out test samples, and reproduces the primary multinomial comparison, directional analyses, qualification comparator, survey-design intervals, ridge and nonlinear checks, cross-fitting, response-style and six-item-domain checks, age and occupation controls, permutation calibration, and leave-one-block-out importance analysis. Respondent-level OECD public-use files and the locally harmonised respondent file are not redistributed. Users must obtain current country files directly from OECD and configure local restricted paths. The archive includes a fast aggregate-only rebuild for tables, Figure 1 and programmed validation.