Work Ledger Generator

Problem

Office documents often sit in scattered folders with inconsistent names, dates, summaries, and references. Manually building a traceable Excel ledger from minutes, proposals, checklists, and task files is error-prone.

How It Works

The skill scans a source directory, creates workbook sheets by subfolder, and derives column structures from actual file content. Core capabilities include:
- Reading .docx, .pptx, .xlsx, .pdf, and legacy .doc files, preferring python-docx, python-pptx, and PyPDF2, then falling back to zipfile plus regex when libraries are unavailable;
- Extracting dates using signature date → filename date → mtime, with support for formats such as YYYY-MM-DD and YYYY年M月D日;
- Pulling 50–200 character summaries from the body text and cleaning XML fragments, Word field-code noise, and split-character spacing;
- Generating blue underlined HYPERLINK formulas so each row can jump back to the source file;
- Supporting incremental updates by detecting dead links and orphan files, completing placeholder summaries, sorting by date, and maintaining version notes.

Boundaries

The skill is intended for readable office documents. Scanned images, pictures, and CAD files are typically labeled by type rather than summarized. If the workbook is open in Excel, close it or save under a non-conflicting name.