prep-slides-master

name

prep-slides-master

description

Re-slim and brand-fix the Equal Experts PRESENTATION / SLIDE-DECK master template (the .pptx behind kuat-studio) when the raw Google Slides master changes. Slides-only. Use when a new "2026 EE branded master slides" export needs turning into the shipped assets/slides/ee-master-2026.pptx — strips example slides + orphan media, sets the theme font to Lexend, embeds Lora + JetBrains Mono, removes Montserrat, and self-checks. Contributor-only; repo-local. NOT for building a deck (that's create-presentation), and NOT for web/marketing/imagery.

Prep the studio slide-deck master

Slides only. This skill maintains the presentation master (.pptx) that the kuat-studio slide skills build from. It has nothing to do with web, marketing, or imagery.

The studio asset pack ships one genuine master, assets/slides/ee-master-2026.pptx, that create-presentation builds from and review-presentation compares against. When the canonical Google Slides master changes, regenerate the shipped master from the fresh raw export — never hand-edit the committed .pptx.

The raw 51 MB export stays out of git by design (provenance in docs/migration/LOG.md, Phase 4S decision A). You need it on hand to run this skill.

Step 1 — Get the raw master

Obtain the latest 2026 EE branded master slides.pptx (PPTX export of the canonical Google Slides master). The script defaults to ~/Downloads/2026 EE branded master slides.pptx; pass another path as the argument.

Step 2 — Run the prep script

The bundled assets/slides/prep-master.py operates directly on the OOXML package (zip + targeted XML edits), not via python-pptx — python-pptx doesn't model embeddedFontLst and a round-trip can silently drop the embedded fonts that are the whole point. It is pure stdlib (no install needed).

python3 assets/slides/prep-master.py "/path/to/2026 EE branded master slides.pptx"

It performs, then self-checks:

1. Slim — drop the example slides + their notesSlides, then prune orphan media (anything no retained master/layout/theme part still references).
2. Fonts — theme font Arial → Lexend; strip off-brand Montserrat; embed Lora + JetBrains Mono (from the committed OFL TTFs in assets/slides/fonts/); sweep stray Arial/Montserrat run typefaces in retained layouts/master → Lexend.
3. Check — assert a valid package with the expected font state and no dangling font/media rels; it refuses to write a broken file.

Expected output (reference run): 51.5 MB → 18.9 MB, slides 79 → 0, layouts kept: 65, embedded: JetBrains Mono, Lexend (+Light/Medium/SemiBold), Lora, self-check: OK.

Step 3 — Verify against the manifest, then commit

• Confirm the self-check passed and the counts match assets/slides/assets.manifest.json: layoutCount: 65, themeFont: Lexend, embeddedFonts includes Lexend/Lora/JetBrains Mono.
• If layout indices changed, re-run curate-slide-layouts and update the manifest's verified layouts (title/section/content/blank) — create-presentation depends on them.
• Sanity-check the size against the verifier's guard: the master must stay ≤ 30 MB (verify-plugins.mjs asset-pack check). If it grew, prune layouts (curate-slide-layouts) before committing.
• Rebuild the studio payload and verify:

  npm run build:plugins && npm run verify:plugins   # asset master present (…MB), manifest resolves
  

Step 4 — Update fonts (only if the brand font set changes)

The non-embedded brand fonts (Lora, JetBrains Mono) are committed as static OFL TTFs in assets/slides/fonts/ (instanced from the variable fonts). If a new weight/style is needed, add the TTF there and extend NEW_FONTS in prep-master.py. Keep the OFL-*.txt licences alongside.

Related

• Part of the contribution model — this is the Medium · slide-asset path (proposing a slide asset).
• Layout map: curate-slide-layouts · brand assets: add-brand-asset
• Consumers: create-presentation, review-presentation
• Script: assets/slides/prep-master.py · manifest: assets/slides/assets.manifest.json