Web Asset Generator: Favicons, Icons, and Social Images from Text
Every web project needs favicons, app icons, and social media images. Designers charge hundreds. Online generators produce generic results. This skill generates production-ready assets through conversation.
Tell Claude what you want. It creates the files, validates them, and integrates them into your project.
What It Creates
- Favicons: Full set from 16x16 to 512x512, all formats
- App icons: iOS, Android, and PWA-ready icons
- PWA manifests: Complete manifest.json with all required fields
- Social images: Facebook, Twitter/X, and LinkedIn preview images
How to Install
Clone the repo and copy the skill:
git clone https://github.com/alonw0/web-asset-generator.git ~/.claude/skills/web-asset-generator
Requires Python with Pillow installed (pip install Pillow).
What Makes It Good
60+ curated emoji options with intelligent suggestions based on your project type. WCAG contrast compliance built in — your assets will pass accessibility checks automatically. Framework detection identifies whether you are using Next.js, React, or vanilla HTML and offers to inject the assets into the right files.
Use it early in your project. Run this skill right after scaffolding a new app. It detects your framework, generates every asset you need, and wires them into your HTML/manifest. You skip 30 minutes of manual favicon setup.
Key Takeaway Professional web assets in 60 seconds. No Figma, no Photoshop, no favicon generators. Describe your brand, pick an emoji, and Claude handles the rest.
GitHub: github.com/alonw0/web-asset-generator
Jobs-to-Be-Done: The Framework Netflix and Intuit Use
Clayton Christensen's Jobs-to-Be-Done theory changed how the best companies build products. The core idea: customers do not buy products. They hire them to make progress in a specific situation.
This skill teaches Claude the full JTBD framework and applies it to your product decisions in real time.
What It Does
- Customer discovery: Generates interview questions that reveal what job your customers are hiring your product to do
- Churn diagnosis: Analyzes why customers leave through the JTBD lens — they found a better way to do the job
- Competitive analysis: Maps competitors not by feature lists, but by which jobs they serve
- Product positioning: Frames your product around the customer's desired outcome, not your feature set
- Three dimensions: Analyzes functional jobs (what they need done), emotional jobs (how they want to feel), and social jobs (how they want to be perceived)
How to Install
Clone the skills repo and copy the JTBD skill:
git clone https://github.com/wondelai/skills.git
Then copy skills/jobs-to-be-done/ to ~/.claude/skills/jobs-to-be-done/
Start with one question. Tell Claude: "I'm building [your product]. Use the JTBD framework to identify the primary job my customers are hiring this product to do." The answer will reshape how you think about your entire roadmap.
Key Takeaway JTBD is the difference between building features nobody asked for and building the thing your customer was going to pay for anyway. This skill gives you a product strategist on demand.
GitHub: github.com/wondelai/skills
Humanizer: Kill the AI Voice
Everyone can spot AI writing. The filler phrases, the em dashes, the "testament to" and "landscape of" language. This skill identifies and removes 25 distinct AI writing patterns based on Wikipedia's documented "Signs of AI writing."
What It Catches
- Content issues: Significance inflation, vague attributions, superficial analysis that sounds deep but says nothing
- Language problems: AI vocabulary ("testament," "landscape," "delve"), copula avoidance, synonym cycling where Claude rotates through fancy words
- Style flaws: Em dash overuse, excessive boldface, title case headings where they do not belong, unnecessary emoji
- Communication artifacts: Chatbot phrases like "I hope this helps," sycophantic tone, over-qualification
- Filler and hedging: Wordy phrases, excessive qualifiers, "it's worth noting that" preambles
How to Install
Clone directly to your skills folder:
git clone https://github.com/blader/humanizer.git ~/.claude/skills/humanizer
How It Works
The Humanizer runs a two-pass review. First pass: rewrite the text to remove detected patterns. Second pass: audit the rewrite for anything that still reads as "AI generated." If a sentence would make someone think "a chatbot wrote this," it gets flagged and rewritten again.
Use it as a final step. Write your content with Claude first. Then run Humanizer on the output. The before/after difference is dramatic — same information, completely different voice.
Key Takeaway AI-generated content that reads like AI-generated content defeats the purpose. Humanizer is the difference between "Claude wrote this" and "a person wrote this." Install it. Use it on everything.
GitHub: github.com/blader/humanizer
Code Explainer: Understand Any Codebase in Minutes
You open a new project. Thousands of files. Unfamiliar patterns. You have no idea where to start. Code Explainer breaks down any code into plain English so you can understand what it does without reading every line.
What It Does
- Function-by-function breakdown: Explain what each function does, why it exists, and how it connects to the rest of the codebase
- Architecture overview: Map the high-level structure of any project — data flow, component relationships, API boundaries
- Pattern recognition: Identify design patterns, conventions, and idioms used in the code
- Beginner-friendly language: Explanations calibrated for your experience level — no jargon, no assumed knowledge
How to Install
Available on the Smithery marketplace:
Visit smithery.ai/skills/CuriousLearner/code-explainer and follow the one-click install.
Use it when you inherit a project. Point Code Explainer at the main entry file and ask: "Explain the architecture of this project and how the key components connect." You will understand the codebase in 5 minutes instead of 5 hours.
Key Takeaway Reading code is the bottleneck for every beginner. Code Explainer removes that bottleneck. You learn faster, contribute sooner, and stop feeling lost in unfamiliar projects.
Smithery: smithery.ai/skills/CuriousLearner/code-explainer
Install All 5 in Under 5 Minutes
Open your terminal. Run these commands in order.
1
Marketing Skills
npx skills add coreyhaines31/marketingskills
2
Web Asset Generator
git clone https://github.com/alonw0/web-asset-generator.git ~/.claude/skills/web-asset-generator
3
Jobs-to-Be-Done
git clone https://github.com/wondelai/skills.git /tmp/wondelai-skills && cp -r /tmp/wondelai-skills/jobs-to-be-done ~/.claude/skills/jobs-to-be-done
4
Humanizer
git clone https://github.com/blader/humanizer.git ~/.claude/skills/humanizer
5
Code Explainer
Visit smithery.ai/skills/CuriousLearner/code-explainer and click install.
First Tasks to Try
- Marketing Skills: "Audit my landing page at [URL] and give me 5 conversion improvements"
- Web Asset Generator: "Generate a full favicon set for my project using a rocket emoji on a dark blue background"
- Jobs-to-Be-Done: "I'm building [product]. Identify the primary job my customers hire this product to do"
- Humanizer: Paste any AI-generated text and say "Humanize this"
- Code Explainer: "Explain the architecture of this project and how the main components connect"
Key Takeaway Five skills. Five minutes. Claude goes from a general assistant to a marketing team, design tool, product strategist, writing editor, and code tutor. All free.