# Slides Grab > End-to-end presentation workflow usable in Codex and Claude Code. Routes the user request to the HTML-mode or image-native-mode sub-skill, then runs the shared plan, design, and export stages. PDF and per-slide PNG are preferred; PPTX/Figma export is experimental / unstable. - Skill: `nota-america/slides-grab` (Agent Skill, multi-file: 2 files) - Install (CLI): `npx skillmds@latest add nota-america/slides-grab` - Raw SKILL.md: https://api.skillmd.com/api/skills/nota-america/slides-grab/raw - Safety review: pending - Works with: Claude Code, Claude.ai, OpenAI Codex - Category: Design & Media - Author: nota-america (https://skillmd.com/u/nota-america) - Updated: 2026-09-17 - Page: https://skillmd.com/skills/nota-america/slides-grab --- # slides-grab Skill - Mode Router Decides how the deck should be built and routes to the right pipeline. Read the user's intent, pick a delivery mode, then hand off to the matching sub-skill for the detailed stage instructions. Both pipelines share the same Stage 1 (`slides-grab-plan`) and Stage 3 (`slides-grab-export`) skills; the difference is how Stage 2 produces slides. --- ## Pick a delivery mode Ask the user (or infer from the request) which pipeline to use: | Signal | Route to | | --- | --- | | "make slides about X", "write a deck", editable text, searchable content, accessibility, charts, diagrams, Card-news, most greenfield decks | **slides-grab-html** — semantic HTML slides | | "make slides in this template/form", "match this existing deck", visual fidelity to a corporate template matters more than editable text, user handed a filled PPTX/PDF and wants the same look | **slides-grab-image** — image-native raster wrapper slides | When in doubt, default to **slides-grab-html**. Only choose image-native when the user explicitly prioritizes visual matching to an existing template over editability. After picking the mode, **read the matching sub-skill in full** and follow its pipeline. Do not mix instructions from the two sub-skills — the image generation steps are different and must not be combined. --- ## Shared stages (both pipelines) ### Stage 1 — Plan Use the installed **slides-grab-plan** skill. The plan stage records the chosen `mode: html` or `mode: image-native` in `slide-outline.md` meta so Stage 2 knows which pipeline to run. Do not proceed to Stage 2 without approval of both style and outline. ### Stage 3 — Export Use the installed **slides-grab-export** skill. Export is gated on a Stage 2 design-gate `Proceed` verdict. PPTX and Figma are experimental / unstable. 1. Confirm user wants conversion. 2. Pick the primary target: - Card-news / Instagram-style decks → `slides-grab png --slides-dir --slide-mode card-news --resolution 2160p` (see `slides-grab-card-news`). - Widescreen decks → `slides-grab pdf --slides-dir --output .pdf`. 3. Per-slide PNG (any mode): `slides-grab png --slides-dir --output-dir /out-png --resolution 2160p`. 4. PPTX (optional, **experimental / unstable**): use `slides-grab convert --slides-dir --output .pptx --engine raster` for visual fidelity (default), or `--engine text` for best-effort editable text. 5. Figma-importable PPTX (optional, **experimental / unstable**): `slides-grab figma --slides-dir --output -figma.pptx`. 6. Report results. --- ## Shared rules (apply to both pipelines) 1. **Always follow the stage order**: Plan → Design → Export. 2. **Get explicit user approval** before advancing to the next stage. 3. **Pass the design gate before export**: Stage 2 must end with a `Proceed` verdict (zero unresolved Critical findings) on the latest rendered state. `slides-grab pdf`, `slides-grab convert`, and `slides-grab figma` block if the receipt is missing, stale, or fails the CLI-enforced report contract. 4. **Use `decks//`** as the slides workspace for multi-deck projects. 5. **Call out export risk clearly**: PPTX and Figma export are experimental / unstable. 6. For complex diagrams, default to a `tldraw`-generated asset unless the user explicitly asks for a different approach. 7. When a slide needs bespoke imagery inside an HTML deck, prefer the default codex-imagen provider via `slides-grab image` (reuses local Codex ChatGPT login — no API key required; run `codex login` once) and keep the saved asset local under `/assets/`. Image-native decks generate each whole-slide raster with `slides-grab image --image-native --name slide-XX` so the wrapper and editor metadata are created together — see the image sub-skill. 8. When a slide needs a chart, default to Chart.js in-slide canvas rendering and require `slides-grab validate` plus `slides-grab build-viewer` review so blank canvases are caught as `empty-canvas` before export. Local videos live under `/assets/` with `poster="./assets/"` thumbnails; use `slides-grab fetch-video` or `yt-dlp` to pull supported web videos first. 9. Use `slides-grab list-styles` and `slides-grab preview-styles` for bundled style discovery during planning. --- ## Reference - `references/presentation-workflow-reference.md` — archived end-to-end workflow guidance from the legacy skill set - `../slides-grab-html/SKILL.md` — HTML-mode pipeline (semantic slides) - `../slides-grab-image/SKILL.md` — image-native pipeline (raster wrapper slides)