Deck Builder logo

Deck Builder

Community
rodrigolourencofarinha
deck-builder

Plan, critique, redesign, and produce slide decks from deck.md briefs. Use for consulting-style decks, executive reports, MBA/class decks, workshops, talk decks, designer-mode PDF output with gpt-image-2, editable ppt-shapes PPTX output with artifact-tool, and visual redesigns from user-supplied decks, templates, screenshots, logos, brand guides, icon packs, or reference assets.

Overview

Publisherrodrigolourencofarinha
Repositorydeck.md
Skill namedeck-builder
Stars
4
Forks
0
Bundled files
32
LicenseMIT
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • 32 bundled files

    Scripts, templates, and references the model can read while it works. Files are read-only and never executed.

  • Open source

    Published by rodrigolourencofarinha on GitHub. Read the source before you install it.

Installation

Install the Deck Builder AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/rodrigolourencofarinha/deck.md/tree/main/skill
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/rodrigolourencofarinha/deck.md.git /tmp/deck.md
mkdir -p .claude/skills
cp -r /tmp/deck.md/skill .claude/skills/deck-builder
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Deck Builder in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Deck Builder on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Deck Builder is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

Deck Builder

Use this skill to turn messy material into a reviewable deck.md, get the human to approve it, and then produce slides from that approved brief.

Core Contract

deck.md is the planning artifact and source of truth. Do not invent a parallel schema.

Hard rules:

  • Draft or normalize a complete deck.md first.
  • Keep status: draft until the human approves production.
  • Do not build slides, generate images, assemble PDFs, or export PPTX files before approval.
  • Always include a first slide with type: cover; it introduces the deck and does not replace the executive summary.
  • Default production_defaults.default_slide_mode to designer-mode unless the user asks for ppt-shapes or mixed/editable output.
  • Declare every model input in designer_assets before approval; do not send or use undeclared files.
  • Treat all logos, templates, old decks, screenshots, brand guides, icon packs, fonts, and reference images as external assets supplied by the user or declared by URL/path. This public skill has no bundled visual asset library.
  • If the request includes data, CSV/Excel, SQL, metrics, benchmarks, or asks for analysis/rationale, complete the data-analysis workflow before drafting deck.md.
  • Make the takeaway unmistakable on every slide. Use charts as proof, not decoration or a catalog of analysis cuts.
  • For pure designer-mode, deliver PDF output only; add OCR/searchable text when available and report if OCR was not applied.
  • Use the standard footer unless the approved brief disables it: small CR lower-left and simple page numbers (1, 2, 3) lower-right.
  • Render and inspect the produced artifact before delivery; repair and rerender failed slides.
  • For post-render changes, create a review deck.md version and regenerate only changed slides.
  • Save meaningful specs and production rounds under the standard instance structure.

Fast Workflow

  1. Classify the request: outline, content plan, data-to-deck, critique, designer-mode PDF, editable PPTX, or mixed deck.
  2. If data or SQL is involved, preserve inputs, create analysis/chart CSVs, write notes and a manifest, then reference them in deck.md.
  3. Draft or update deck.md with the narrative, action titles, slide types, production defaults, declared assets, and analysis artifacts when relevant.
  4. Send the full draft deck.md to the human for approval.
  5. After approval, validate the brief, prepare declared assets, validate data refs, and produce only in the approved mode.
  6. Render, inspect, repair, and save the output in a production instance.
  7. For review changes, fork from the previous approved brief, add ## Revision Brief, and reuse unchanged slides.

If the build is simple, keep the workflow simple. Do not create extra planning files unless they materially help review or production.

Reference Router

Read only the files needed for the task:

  • references/SPEC.md - authoritative deck.md schema, defaults, precedence, and output rules.
  • references/deck.md - quick-start starter brief.
  • references/deck.full.md - rich starter with assets, revision brief, design tokens, and advanced designer-mode controls.
  • references/standards/deck-validation.md - hard validation rules and self-check checklist.
  • references/standards/data-analysis-workflow.md - SQL/CSV/spreadsheet analysis preflight, artifact layout, manifest schema, and chart economy rules.
  • references/standards/slide-archetypes.md - valid slide type values and preferred production modes.
  • references/standards/narrative-templates.md - SCR, pyramid, problem-solution, and update templates.
  • references/standards/artifact-structure.md - specs, assets, instances, image folders, method files, manifests, and outputs.
  • references/standards/image-prompts.md - designer-mode prompt rules and image-generation defaults.
  • references/deck-workflow.md - deck-level storyline and consulting narrative rules.
  • references/consulting-slide-standards.md - action-title, proof, source, and pre-flight standards
  • references/standards/slide-design-principles.md - visual execution: canvas, margins, layout zones, typography hierarchy, whitespace, chart design, color discipline, alignment, and pre-design checklist.
  • references/single-slide-workflow.md - one-slide build or critique sequence.
  • references/designer-mode-workflow.md - designer-mode production and review loop.
  • references/designer-mode-gpt-image-prompt-scaffold.md - convert an approved slide block into a gpt-image-2 prompt.
  • references/designer-mode-direct-api-pattern.md - direct OpenAI Image API fallback, only after Codex path failure and human approval.
  • references/pptx-production.md - editable PPTX production with artifact-tool.

Production Modes

Use designer-mode for premium visual slides and image-backed explanatory pages. Generate slide by slide with gpt-image-2 through Codex OAuth/Codex image tooling first. Use direct API-key generation only after that path fails and the human approves the fallback.

Use ppt-shapes for editable analytical slides, precise charts, tables, process flows, matrices, and decks where the user needs a PowerPoint-native file. Prefer:

bash
python scripts/build_pptx_artifact_tool.py OUTPUT.pptx APPROVED_DECK.md

This is the public editable-PPTX path. It creates an isolated artifact-tool workspace with src/, scratch/, and output/; uses editable presentation primitives; exports PPTX, PNG previews, layout JSON, and a quality report; and prepares only assets declared in the approved brief.

Use mixed only when specific slides need different modes. Set mode: explicitly per slide.

External Asset Contract

The installable skill is intentionally asset-neutral. Do not assume any bundled templates, logo sets, icon packs, deck examples, fonts, or brand materials exist inside the skill.

When the user provides assets, add them to designer_assets before approval. Supported generic types include logo, ppt-template, existing-deck, slide-preview, brand-guide, reference-image, screenshot, icon-pack, font, and other.

For designer-mode, prepare declared non-image assets into model-readable inputs and record them in method/model-inputs.yaml.

For ppt-shapes, pass declared local assets through the artifact-tool workspace under scratch/assets/ and record them in scratch/external-assets.json. Use them only when referenced by scope: deck or slide-level asset_refs.

If an approved deck marks an asset as required: true and it cannot be found or prepared, stop and report the missing asset. Optional missing assets should not block storyline planning or validation.

Validation Before Delivery

Before delivering, check:

  • the current deck.md is approved
  • data-driven decks have analysis_artifacts, local chart CSVs, notes/manifest, and clear caveats
  • narrative template and slide types are valid
  • the first slide is type: cover
  • non-cover slide titles are action titles and the title sequence tells the argument
  • each slide has one consulting takeaway and any chart directly proves it
  • quantitative claims have sources
  • all asset_refs resolve to declared designer_assets
  • prepared model inputs are recorded in method/model-inputs.yaml when designer assets are used
  • output has been rendered and inspected for text accuracy, overlap, clipping, logo placement, safe margins, and footer
  • production artifacts are stored in the instance structure
  • review changes use a new review spec and regenerate only changed slides

Full validation lives in references/standards/deck-validation.md.

Bundled files

The model reads these on demand while the skill is loaded. They are exposed as readable files and are never executed.

Frequently asked questions

What does the Deck Builder AI skill do?

Plan, critique, redesign, and produce slide decks from deck.md briefs. Use for consulting-style decks, executive reports, MBA/class decks, workshops, talk decks, designer-mode PDF output with gpt-image-2, editable ppt-shapes PPTX output with artifact-tool, and visual redesigns from user-supplied decks, templates, screenshots, logos, brand guides, icon packs, or reference assets.

Why use Deck Builder on TypingMind?

Because you install it once and use it with any model. Deck Builder is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Deck Builder in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/rodrigolourencofarinha/deck.md/tree/main/skill. TypingMind reads its SKILL.md and bundles its files and installs it as a skill you can enable per chat.

Which AI models can use Deck Builder?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Deck Builder?

As many as you like. As long as a model supports skills, you can use Deck Builder with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Deck Builder AI skill free?

Yes. It is published on GitHub by rodrigolourencofarinha under the MIT license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇