What the OOXML package audit checks
Modern Word, Excel, and PowerPoint files are ZIP packages governed by Open Packaging Conventions. This tool reads the central directory and supported entries, checks CRC and package paths, parses bounded XML, identifies the single root office-document relationship, reconciles content types, resolves internal relationship targets, and lists external targets without requesting them.
Hidden content, names, and review history
For Excel, sheet states distinguish visible, hidden, and very-hidden sheets; defined names are checked for #REF!, missing sheet targets, invalid local scopes, and external-workbook formulas. Word revision elements and tracking settings, Office comments, and hidden PowerPoint slides are surfaced as review evidence. These checks cannot prove that every application-specific hidden object has been found.
Macros, OLE, custom XML, and metadata
VBA project relationships and parts, ActiveX parts, embedded/OLE objects, custom XML, digital-signature evidence, and rights-management evidence are inventoried. Core, extended, and custom properties remain separate so a reviewer can see exactly what may identify an author or organization. The page never executes code or recursively inspects embedded payloads.
Why the scrubber is deliberately narrow
The optional rebuilt copy removes only selected elements from readable core and extended property XML. It does not claim to remove comments, revisions, custom properties, links, macros, embedded objects, custom XML, thumbnails, cached values, or application-specific history. It refuses signed, protected, structurally invalid, or unreadable packages because a rebuild could invalidate trust evidence or conceal damage. Always inspect the downloaded copy in the destination Office application.
Validation boundary
A clean audit does not prove OOXML conformance, safety, privacy, or visual fidelity. This is a relationship-aware package preflight, not the complete ECMA-376 schemas, Microsoft Office Document Inspector, malware analysis, signature verification, IRM decryption, macro analysis, embedded-file inspection, or rendering test. Binary legacy DOC/XLS/PPT and encrypted OOXML wrapped in an OLE container are outside scope.
Frequently asked questions
Are external links opened?
No. The literal relationship and formula targets are shown, but the browser makes no request to them.
Does “macro present” mean the file is malicious?
No. It means package relationships, content types, or paths indicate macro content. The VBA project is not executed or analyzed.
Why are comments and tracked changes kept in the cleaned copy?
Removing review content changes document meaning and can silently discard work. This scrubber only removes the explicitly listed property fields.
Why is the downloaded copy sometimes larger?
The local deterministic builder stores ZIP entries without compression. The Office content is unchanged except for the listed property XML elements.
Authoritative format references
Package and document structure follow ECMA-376 Office Open XML and Microsoft’s Open XML package overview. Excel sheet-state interpretation follows Microsoft’s hidden worksheet guidance. For release decisions, use a copy and Microsoft’s Document Inspector guidance.
Inspect an OpenDocument package · Inspect a legacy OLE/CFB container · Browse Office document tools