Pick a deliverable shape. The right skills and templates follow from the choice.
If two rows look right, pick the lighter one. Graduate when the artefact outgrows it.
Each route is content prep, then render. Skills are reused across routes.
| Step | Skill / template |
|---|
| 1 | brain-dump-cleaner (if raw) |
| 2 | content-to-layout (Flexible Artefact only — specific templates carry their own structure) |
| 2b | layout-designer (if content needs shapes the seven default layouts don’t cover) |
| 3 | An HTML template or the Flexible Artefact |
| Step | Skill / template |
|---|
| 1 | brain-dump-cleaner (if raw) |
| 2 | content-to-layout (always — the pptx skill expects a layout map) |
| 3 | pptx-from-brain-dump |
PPTX needs a .pptx template file with the house style baked in. Per-project design.md informs Tone and Don’ts only — the template owns palette and typography.
| Skill | Input | Output |
|---|
brain-dump-cleaner | Raw brain-dump | Cleaned brain-dump |
content-to-layout | Cleaned brain-dump | Layout map (per-slide layout tags) |
layout-designer | Brain-dump | Bespoke layout specs for shapes the defaults miss |
pptx-from-brain-dump | Cleaned brain-dump + layout map + .pptx template | .pptx file |
- HTML artefact outgrowing single-file? Move to the Presentation Bootstrap; keep the brain-dump.
- App growing a need for secrets or a database? Wrong bootstrap — move to Simple or Full.
- Presentation needs to be a
.pptx for the audience? Run the same brain-dump through the PPTX route. Don’t try to convert HTML to PPTX.