A fixed-scope, evidence-backed review of one research or data-analysis repository preparing for collaboration, publication, release, or maintainer handover. I trace one documented core workflow from a safe supplied or synthetic input to its main output and separate what is verified from what is only declared or still unknown.

Scope: one repository with up to 25,000 lines of first-party Python code; current default branch and CI; clean-install and packaging instructions; dependencies and environment declarations; input/output contracts, provenance, configuration, deterministic replay; tests around core scientific transformations; release, documentation, and handover readiness.

You receive a Markdown audit report and polished PDF, an evidence table, a severity-ranked findings register, a prioritized remediation plan, and one 30-minute handover call or equivalent written Q&A. Delivery is within 5 business days after complete access and inputs. One factual-correction or clarification round is included.

This audit does not include code fixes, refactoring, test implementation, dependency upgrades, cloud deployment, penetration testing, formal privacy review, statistical reanalysis, or validation of scientific conclusions. I prefer least-privilege read-only access and a public, synthetic, or properly de-identified fixture. Do not send credentials, direct identifiers, secrets, or data you are not authorized to share.