Open source · active development

AuditFlow

A lightweight framework for structured, traceable, and reproducible internal audit work.

The problem

Audit work is often scattered across spreadsheets, documents, emails, screenshots, and slide decks. The final conclusion may be correct, but the reasoning behind it gradually disappears and updating calculations becomes expensive.

AuditFlow Philosophy

Time is the most valuable resource. Therefore:

no work for the sake of work
auditors must think before starting tests
work should be easy to understand and trace, avoiding time wasted on clarification
decisions and conclusions should be data-based, and analysis should be simple and fast: an error is fixed in a script in 5 minutes, not in Excel in 3 hours
mechanical work should be minimized

Traceability chain

Risk → Control → Test → Evidence → Result → Observation → Action → Report

How it works

AuditFlow keeps structured decisions in YAML, audit narratives and workpapers in Markdown/Quarto, raw evidence unchanged, and generated outputs separate from source evidence.

Current limitations

it is not a complete corporate internal audit methodology
basic Python practice is required to get the most from it
deeper methodology validation is still evolving
automated LLM integration is not finished
confidential data should not be sent to an external model without a separate risk assessment