Humanize PPT

Use this skill when a user wants to turn raw material, notes, voice transcripts, documents, links, or old PPTs into a presentation-ready outline before rendering slides.

Positioning

Humanize PPT is an Outline Director and Agent Teams Orchestrator, not a slide renderer.

It should run before downstream PPT / HTML slide skills. Its job is to produce a clean AST-based production brief so renderers do not ingest raw noisy material directly.

In Agent Teams mode, the main Humanize PPT Agent loads this skill and controls specialist agents:

• Guizang Agent for Chinese stable rendering.
• Zara Agent for style exploration, HTML production, and deploy.
• HyperFrames Agent for video slots.
• Presenter Agent for presenter mode after the deck is finalized.
• QA Agent for content, visual, path, and delivery checks.

AST theory

AST means Audience-State-Transfer.

• Audience: who is listening, what they know, what they resist, and why they would keep listening.
• State: the audience state before and after the deck, plus the core tension that blocks the transition.
• Transfer: the slide-by-slide path that moves the audience from initial state to desired state.

Core sentence:

PPT is not an information container. PPT is an audience state-transfer artifact.

Required output contract

For every Humanize PPT run, produce:

1. deck_brief.md — audience, goal, tension, success criteria.
2. ast_outline.md — AST map and narrative arc.
3. slide_plan.json — slide-by-slide plan.
4. speaker_intent.md — what the speaker should do on each slide.
5. asset_manifest.md — screenshots, charts, images, video needs.
6. video_slots.json — optional HyperFrames / video insertion plan.

Recommended OPC workflow

O — Outline Director
  Humanize PPT: raw material → AST outline + production brief

P — Presentation Production
  guizang path: Chinese stable HTML PPT
  Zara path: style exploration and HTML production

C — Complete / Control
  HyperFrames video adapter
  Presenter Adapter shell
  Deploy / export adapter
  QA checklist

Rules

1. Do not let slide renderers consume raw material directly when Humanize PPT can first produce the AST contract.
2. Keep presenter mode as a post-processing adapter, not a style.
3. Separate deployment from presenter mode.
4. Absorb AI-writing cleanup principles from humanizer tools, but do not reduce Humanize PPT to text polishing.
5. Prefer a small verified workflow over a broad unverified promise.
6. For public Skill releases, create/push the repo, install from GitHub locally, run one safe full sample, verify style exploration + presenter mode + deploy URL, and only then polish README details.
7. For Agent Teams development, emit router_plan.json, run_manifest.json, bounded commands/*.md, and separate outputs/<agent>/ directories before wiring real downstream Skills.

Operational references

• references/agent-teams-public-preview.md — Agent Teams architecture, specialist-agent command protocol, public preview release loop, and README split convention.

Local demo

If this repository is installed locally, run:

python3 scripts/humanize_ppt_v1.py \
  --source examples/01-ai-tool-update/source.md \
  --out .humanize-ppt-runs/ai-tool-update \
  --title "AI 工具更新，不只是功能清单"