slides-to-site

name	slides-to-site
description	Convert a Google Slides presentation into a psd401.ai presentation page. Takes a Google Slides URL, reads all slide content, asks for metadata, and generates the markdown file. Use when adding presentations to the psd401.ai website. Triggers on: add presentation to site, slides to site, publish presentation, psd401.ai presentation.
argument-hint	<google-slides-url>
model	claude-opus-5
effort	high
allowed-tools	["Bash","Read","Write","Edit","Glob","Grep"]
extended-thinking	true

Slides to Site

Convert a Google Slides presentation into a psd401.ai presentation page. Reads the full presentation via gws slides, collects metadata from the user, and generates a properly formatted markdown file.

Constants

• psd401.ai repo: /Users/hagelk/non-ic-code/psd401.ai
• Presentations dir: /Users/hagelk/non-ic-code/psd401.ai/src/content/presentations/
• Thumbnails dir: /Users/hagelk/non-ic-code/psd401.ai/public/images/thumbnails/

Workflow

Process each presentation URL the user provides. After each one, ask if they have more. When done, commit and push.

Step 1: Parse the Google Slides URL

Extract the presentation ID from the user-provided URL argument ($ARGUMENTS).

Supported URL formats:

• https://docs.google.com/presentation/d/PRESENTATION_ID/edit → extract PRESENTATION_ID
• https://docs.google.com/presentation/d/PRESENTATION_ID/pub → extract PRESENTATION_ID
• https://docs.google.com/presentation/d/e/2PACX-.../pub → this is a published URL; extract the 2PACX-... portion after /e/
• Just a bare presentation ID

Validate the extracted ID is non-empty. If no URL was provided, ask the user for one.

Step 2: Read the Presentation Content

Run:

gws slides presentations get --params '{"presentationId": "PRESENTATION_ID"}' --format json

Parse the JSON response to extract:

• Title: from title field at the top level
• All slide text: walk slides[].pageElements[].shape.text.textElements[].textRun.content and concatenate
• Speaker notes: from slides[].slideProperties.notesPage.pageElements[].shape.text.textElements[].textRun.content

If gws fails with an auth error, tell the user:

Auth issue detected. Run: gws auth login -s slides,drive
Then try again.

Step 3: Ask User for Metadata

Use AskUserQuestion to collect all metadata in a single prompt. Show the extracted title as a default. Ask for:

I extracted the presentation content. Here's what I need from you:

**Title**: {extracted_title} (press Enter to keep, or type a new one)
**Date** the presentation was/will be given (YYYY-MM-DD):
**Presenters** (comma-separated names):
**Audience** (e.g., "PSD Community", "PSD Staff", "WTA Conference", "WSSDA Conference"):
**Type** (e.g., "School Board Presentation", "Public Workshop", "PD Session", "Conference Session"):
**Description** (1-2 sentences — or type "auto" and I'll generate one from the content):
**Slug** (kebab-case filename, suggested: {auto_slug}) — press Enter to accept or type a new one:

Where {auto_slug} is generated by lowercasing the title, removing non-alphanumeric chars, and converting spaces to hyphens.

If the user says "auto" for description, synthesize a 1-2 sentence summary from the slide content.

Step 4: Generate the Embed URL

Convert the input URL to embed format:

• For standard URLs (/d/PRESENTATION_ID/): https://docs.google.com/presentation/d/PRESENTATION_ID/embed
• For published URLs (/d/e/2PACX-.../): https://docs.google.com/presentation/d/e/2PACX-.../embed

Step 5: Export Thumbnail

Auto-generate the thumbnail using the Slides API thumbnail endpoint:

1. Extract the first slide's objectId from the presentation JSON fetched in Step 2 (it's slides[0].objectId)
2. Call the thumbnail API:

gws slides presentations pages getThumbnail --params '{"presentationId": "PRESENTATION_ID", "pageObjectId": "FIRST_SLIDE_OBJECT_ID"}' --format json

3. Parse the contentUrl from the JSON response
4. Download it with curl:

curl -sL -o /Users/hagelk/non-ic-code/psd401.ai/public/images/thumbnails/{slug}.png "CONTENT_URL"

5. Verify the file was created and is a valid PNG using file command

If any step fails (permissions, auth, etc.), inform the user:

Could not auto-generate thumbnail. Please manually save the first slide as:
/Users/hagelk/non-ic-code/psd401.ai/public/images/thumbnails/{slug}.png

Set the thumbnail path in frontmatter regardless: /images/thumbnails/{slug}.png

Step 6: Generate Markdown Content

Using the full slide text content, generate the markdown body. Follow the exact format used by existing presentations in the repo.

Structure:

1. Bold title — repeats the frontmatter title
2. Context paragraph — who presented, when, where, to whom, and a summary of what the presentation covers
3. Key Takeaways — 4-7 bullet points with bold headers followed by details, extracted/synthesized from actual slide content
4. Actionable Insights — 3-5 bullet points with bold headers and practical recommendations
5. Looking Ahead — 1 paragraph on future directions or implications

Step 7: Write the File

Assemble the complete markdown file using this exact template:

---
title: '{title}'
date: '{YYYY-MM-DD}'
presenters:
  - '{Presenter 1}'
  - '{Presenter 2}'
audience: '{audience}'
type: '{type}'
thumbnail: '/images/thumbnails/{slug}.png'
slides: '{embed_url}'
description: '{description}'
---

**{Title}**

{Context paragraph — who presented this, when, where, to what audience, and what it covers.}

**Key Takeaways:**

- **{Point 1 Header}:** {Details synthesized from slide content}
- **{Point 2 Header}:** {Details synthesized from slide content}
- **{Point 3 Header}:** {Details synthesized from slide content}
- **{Point 4 Header}:** {Details synthesized from slide content}

**Actionable Insights:**

- **{Insight 1}:** {Practical recommendation}
- **{Insight 2}:** {Practical recommendation}
- **{Insight 3}:** {Practical recommendation}

**Looking Ahead:**

{Future directions paragraph — what comes next, implications, or next steps.}

Critical formatting rules:

• Frontmatter values with colons or special chars MUST be wrapped in single quotes
• Each presenter gets its own - 'Name' line under presenters:
• No tags: field — the site derives tags from presenters, type, and audience
• One blank line between each section
• Use **bold** for section headers and bullet point leaders

Write the file to: /Users/hagelk/non-ic-code/psd401.ai/src/content/presentations/{slug}.md

After writing, read the file back to verify it was created correctly and frontmatter is valid.

Show the user:

✓ Created: src/content/presentations/{slug}.md
  Title: {title}
  Presenters: {presenters}
  Slides embed: {embed_url}

Step 8: Ask for More Presentations

Ask the user:

Do you have more presentations to add, or should I commit and push?

• If they have more → loop back to Step 1 with the next URL
• If done → proceed to Step 9

Keep track of all slugs created in this session for the commit.

Step 9: Commit and Push to Deploy

Once the user confirms they're done:

cd /Users/hagelk/non-ic-code/psd401.ai
git add src/content/presentations/{all-slugs}.md public/images/thumbnails/{all-slugs}.png
git commit -m "content: add presentation(s) — {title(s)}"
git push origin main

After pushing, inform the user:

Pushed to main. AWS Amplify will auto-deploy the changes.
New presentation(s) will be live at psd401.ai/presentations shortly.

If any thumbnail files don't exist (manual export needed), only git add the markdown files and remind the user to add thumbnails later.