ZKT PPT Beautification Standard Template

Problem

Raw data PPTs often contain tables and screenshots but lack a consistent cover, table of contents, KPI highlights, and editable charts. Applying a template manually can mix source images into template positions or produce chart images that cannot be edited. This skill extracts titles, tables, KPI values, trend series, and images from a source file, then maps them to a fixed visual specification.

How It Works

Core workflow:
- Parse source .pptx: extract body text, tables, KPI metrics, trend data, and images.
- Generate or load the standard template: template assets are auto-downloaded from IMAGE_URLS, avoiding source PPT covers or backgrounds.
- Rebuild content slides: add TOC, section dividers, native editable charts (bar, line, doughnut), styled tables, and data cards.
- Run overlap checks: detect bounding-box collisions, adjust spacing, and regenerate until all slides pass.

Style constraints include a limited palette, Microsoft YaHei and Arial Black typography, and flat shapes with no shadows or strokes.

Boundaries

Before generation, the workflow asks whether the user is a ZKT employee and whether to beautify strictly from source content or use the default generation logic. Strict source mode keeps content 100% from the source file while still applying cover, header, logo, and color rules. Default mode adds company pages, TOC, and section dividers. If the source lacks clear chart series or KPI metrics, output may remain table-heavy. Python 3 is required, and the overlap check must pass before delivery.