AgriciDaniel/claude-repurpose / skills/repurpose
repurpose
VerifiedContent Repurposing Engine. Transforms any content (YouTube videos, blog posts, podcasts, local files, pasted text) into platform-optimized outputs for Twitter/X, Threads, LinkedIn, Instagram, TikTok, Pinterest, Snapchat, Facebook, YouTube Community, Skool, Discord, Reddit, Quora, Medium, WhatsApp Channels, Telegram, and email newsletters. Atomizes content into reusable pieces, adapts brand voice per platform, generates polls, image prompts (via /banana), publishing calendars, and SEO metadata. Triggers on: "repurpose", "content repurpose", "turn into social media", "cross-platform", "content atomization", "repurpose blog", "repurpose video", "turn this into posts", "threads", "pinterest", "snapchat", "discord", "medium", "whatsapp", "telegram".
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 19.8 KB |
| references/engagement-benchmarks.md | markdown | 7.2 KB |
| references/hook-formulas.md | markdown | 9.5 KB |
| references/image-sourcing.md | markdown | 7.0 KB |
| references/mistakes-to-avoid.md | markdown | 4.0 KB |
| references/platform-specs.md | markdown | 7.3 KB |
| references/poll-strategy.md | markdown | 4.2 KB |
| references/repurposing-frameworks.md | markdown | 8.1 KB |
| references/voice-adaptation.md | markdown | 8.5 KB |
references/image-sourcing.md
View on GitHub# Image Sourcing Reference
Comprehensive guide for sourcing, generating, and sizing images during content repurposing.
Replaces `image-prompts.md`. Three tiers of image sourcing, prioritized by relevance.
---
## 3-Tier Image Priority Chain
### Tier 1: Website Images (when input is a URL)
`extract_article.py` already returns an `images` array with `src` + `alt` for every image
on the source page. These are the MOST relevant images because they came from the source
content itself.
**Filtering rules** (applied by `scripts/fetch_images.py --article-images`):
- Skip images with "logo", "icon", "avatar", "favicon", "sprite", "banner-ad",
"advertisement" in URL or alt text
- Skip images smaller than 400px (detect via URL size hints like `150x150`, `thumbnail`,
`thumb`)
- Skip SVGs (usually icons or vector decorations)
- Rank by keyword overlap between alt text and content atoms
- Return top N images sorted by relevance score
**Command:**
```bash
python3 scripts/fetch_images.py --topic "WordPress AI automation" \
--article-images '{"images": [...]}' --count 5 --verify
```
### Tier 2: Stock Photos (Pixabay -> Unsplash -> Pexels)
Search via the WebSearch tool with platform-specific queries generated by `fetch_images.py`:
| Platform | Query Template | Attribution |
|----------|---------------|-------------|
| Pixabay (preferred) | `site:pixabay.com [topic] wide professional` | None required |
| Unsplash (resize support) | `site:unsplash.com [topic] professional` | Appreciated, not required |
| Pexels (fallback) | `site:pexels.com [topic] high quality` | None required |
**Extract direct CDN URLs from search results:**
- Pixabay: `https://cdn.pixabay.com/photo/YYYY/MM/DD/HH/MM/topic-12345_1280.jpg`
- Unsplash: `https://images.unsplash.com/photo-{id}?w=1200&h=630&fit=crop&q=80`
- Pexels: `https://images.pexels.com/photos/{id}/pexels-photo-{id}.jpeg?auto=compress&w=1200`
**Rules:**
- Verify every URL with a HEAD request before using
- Target 3-5 relevant images per repurpose run
- Prefer Pixabay (no attribution), then Unsplash (resize params via URL), then Pexels
### Tier 3: AI-Generated (Gemini via /banana or direct API)
| Always used for | Used as fallback when | Used for |
|----------------|-----------------------|----------|
| Quote cards (require text overlay) | Tier 1-2 produce < 3 suitable images | Carousel covers (need title text) |
| | Stock search returns < 3 results | Abstract or niche topics |
**Model:** `gemini-3.1-flash-image-preview` (NB2 Flash, ~$0.067/image)
**Cost estimate for typical run:**
- 5 quote cards + 2 carousel covers + 3 hero images = ~10 images
- Estimated cost: $0.40-$0.80 total
- RULE: Show cost estimate before batch generation. Never auto-generate > 3 without confirmation.
---
## 6-Component Image Brief (for AI generation)
Every AI image prompt follows this structure (upgraded from 5-component to match claude-blog):
```
Subject: [Who/what with physical detail -- textures, materials, scale]
Action: [What's happening -- pose, gesture, movement, state]
Context: [Environment -- time of day, weather, setting, atmosphere]
Composition: [Camera -- angle, shot type, framing, depth of field]
Lighting: [Light source -- direction, temperature, shadows, quality]
Style: [Art medium -- film reference, technical specs, aesthetic]
```
### Quote Card Template (AI-only)
```
Subject: A minimalist quote card with modern typography displaying "[QUOTE]"
Action: Clean text prominently centered with subtle decorative accents
Context: [MOOD_COLOR] gradient background, [dark navy/electric blue/warm amber] palette
Composition: Centered layout, generous whitespace, attribution text below quote
Lighting: Soft ambient glow highlighting the text, subtle vignette at edges
Style: Modern social media design, sharp and professional, high contrast text on dark background
```
- One quote per card; never stack multiple quotes
- Keep quote under 120 characters for readability
- Attribution format: "-- [Author Name]" or "-- [Author], [Title/Company]"
- Leave 10% margin on all sides for platform safe zones
### Carousel Cover Slide Template
```
Subject: Bold typographic cover slide reading "[TITLE]"
Action: Eye-catching title treatment with strong visual hierarchy
Context: [BRAND_COLOR] background with subtle pattern or texture
Composition: Title fills 60% of frame, subtitle or byline at bottom
Lighting: Even studio lighting, no harsh shadows on text
Style: Professional presentation slide, clean and modern, high readability
```
### Social Hero Image Template
```
Subject: Conceptual illustration representing [TOPIC]
Action: Visual metaphor for [KEY_CONCEPT]
Context: Clean professional setting with subtle depth
Composition: Balanced composition with clear focal point, text-safe areas on left/right third
Lighting: Soft directional light creating gentle depth, warm temperature
Style: Modern editorial illustration, flat design with depth, brand-consistent colors
```
---
## Platform Dimension Reference
See `platform-specs.md` for the full cross-platform image size table. Key sizes for image sourcing:
| Platform | Pixels | Unsplash Resize Params |
|----------|--------|------------------------|
| Twitter/X | 1600x900 | `?w=1600&h=900&fit=crop&q=80` |
| LinkedIn / Instagram / Facebook | 1080x1350 | `?w=1080&h=1350&fit=crop&q=80` |
| YouTube Community | 1080x1080 | `?w=1080&h=1080&fit=crop&q=80` |
| Newsletter / OG | 1200x630 | `?w=1200&h=630&fit=crop&q=80` |
| Quote card | 1080x1080 | `?w=1080&h=1080&fit=crop&q=80` |
---
## Alt Text Rules
- 10-125 characters
- Full descriptive sentence, not keyword stuffing
- Include topic relevance naturally
- For stock photos: describe what the image shows AND its content connection
- For AI images: describe the visual AND the quote/message
- For charts: include the key data point
**Good:** "WordPress dashboard showing AI-managed blog posts with automatically generated images"
**Bad:** "WordPress AI blog image SEO content marketing"
---
## When to Use Which Tier
| Scenario | Tier |
|----------|------|
| Input is a blog URL with good images | 1 (Website) |
| Input is YouTube video (no images) | 2 (Stock) then 3 (AI) |
| Input is local file / pasted text | 2 (Stock) then 3 (AI) |
| Need quote cards with text overlay | 3 (AI ONLY) |
| Need carousel cover with title text | 3 (AI) |
| Topic is niche or abstract | 3 (AI) |
| Stock search returns < 3 results | 3 (AI fallback) |
---
## Fallback Chain
```
Website images available?
YES --> Use them (Tier 1)
NO |
v
Stock photos found (3+)?
YES --> Use them (Tier 2)
< 3 |
v
/banana available?
YES --> Generate AI images (Tier 3)
NO |
v
Save prompts to banana-prompts.md --> User generates manually
```
### banana-prompts.md Format (when /banana is NOT available)
```markdown
### [Image Name] - [Platform] ([Aspect Ratio])
**Use:** [quote card / cover slide / hero image]
**Prompt:** [Full 6-component prompt]
**Filename:** [suggested-filename.png]
```
---
## /banana Detection
Check for `gemini_generate_image` MCP tool in current session, or `~/.claude/skills/banana/SKILL.md` on disk. If neither is available, save prompts to `banana-prompts.md` for manual generation.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 471k | Updated | Claude CodeCodex | Verified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |