pptx-encoder

PPTX Encoder

Use this skill when the team must render deck_spec.json and style_spec.json into a real editable .pptx.

Tool

Use:

python .opencode/skills/pptx-encoder/scripts/render_deck.py \
  --deck-spec deck_spec/deck_spec.json \
  --style-spec deck_spec/style_spec.json \
  --output delivery/final_deck.pptx \
  --report encoding/encoding_report.md

or from the repository:

python teams/ppt_writer_team/skills/pptx-encoder/scripts/render_deck.py \
  --deck-spec deck_spec/deck_spec.json \
  --style-spec deck_spec/style_spec.json \
  --output delivery/final_deck.pptx \
  --report encoding/encoding_report.md

Supported Elements

text
bullets
callout
table
image_placeholder
image
video_link
chart_placeholder
divider
footer/source note
speaker notes

Style Spec Support

The renderer reads these optional fields from style_spec.json:

typography: title/subtitle/body/caption sizes and minimum readable sizes.
layout_tokens: margins, gutters, header height and footer height.
template_patterns: reusable header/footer behavior and deck-specific header/footer text.
theme_colors: background, text, primary, secondary, accent, muted and surface.
template_pptx: optional path to Boss-provided template. When present, renderer uses it as the base presentation where possible.
template_fidelity: preserve_background/preserve_master/header_footer flags.

The renderer dynamically reduces text size within configured minimums and distributes multiple elements across the available content area. If a slide still needs unreadably small text, revise deck_spec.json by splitting content across slides instead of forcing more text into one slide.

Image Elements

Use prepared local image assets:

{
  "type": "image",
  "image_path": "materials/images/processed/hero-product-1600x900.jpg",
  "caption": "Optional short caption"
}

image_placeholder can also include image_path; if the file exists, the renderer inserts the image. If the path is missing, it renders a placeholder and records a warning.