extract-project-logo
Skill834 starsChanged 3 days agoMITView on GitHub
Extract a project's logo from its website, brand kit, or other sources
In this repository23 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- add-curatorSkill
- add-stablecoinSkill
- add-vault-abi-methodsSkill
- add-vault-noteSkill
- add-vault-protocolSkill
- categorise-vault-strategySkill
- check-stablecoinsSkill
- check-top-list-vaultsSkill
- check-vault-onchainSkill
- clean-up-for-prSkill
- document-curator-incidentSkill
- extract-test-setSkill
- extract-vault-protocol-logoSkill
- find-new-curatorsSkill
- fix-sphinx-docsSkill
- get-block-numberSkill
- Identify vault protocolSkill
- maintain-curatorsSkill
- make-releaseSkill
- monthly-vault-reportSkill
- more-vaultsSkill
- post-process-logoSkill
The file
Show its rules and text
We split this skill into the rules it gives an agent
37 rules. About 0 are common, 31 are distinctive, and 6 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈31
Distinctive · estimate
This author's own choices. Worth a look.
6
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 12
Before starting, gather the following from the user:
- Distinctive: line 18
If any required input is missing, ask the user before proceeding.
- Distinctive: line 47
If a brand kit is found:
- Distinctive: line 48
Look for downloadable logo packages (often ZIP files)
- Distinctive: line 49
Prefer SVG or vector formats over raster images
- Distinctive: line 51
If both light and dark themes are present, get both
- Distinctive: line 52
If colourful option exist, get it as well
- Distinctive: line 54
If a brand kit is found with suitable logos, proceed to Step 5.
- Distinctive: line 62
Search these common locations:
- Distinctive: line 69
Check the README.md for embedded logos:
- Very specific: line 70
Look for <img> tags or markdown image syntax
- Distinctive: line 73
Search for files named:
- Distinctive: line 78
If a suitable logo is found, proceed to Step 4.
- Distinctive: line 82
If no brand kit or GitHub assets are available, extract the logo from the website:
- Distinctive: line 86
Use the MCP Playwright tool to:
- Distinctive: line 89
Look for logo elements in the header (typically top-left):
- Very specific: line 90
Search for <img> tags with class/id containing "logo"
- Very specific: line 91
Check for <svg> elements in the header
- Very specific: line 92
Look for elements with role="img" and logo-related aria labels
- Very specific: line 98
Open Graph image: - Check for <meta property="og:image"> tag
- Distinctive: line 102
Favicon: - Check for high-resolution favicon:
- Very specific: line 107
Twitter Card image: - Check <meta name="twitter:image">
- Distinctive: line 121
Ensure the file is valid XML
- Distinctive: line 126
Verify file integrity after download
- Distinctive: line 131
Check file size is reasonable (not 0 bytes or corrupted)
- Distinctive: line 166
Create a Markdown file in the folder where logos where saved
- Distinctive: line 174
If no suitable logo can be found:
- Distinctive: line 176
Check if the project uses a different primary domain
- Distinctive: line 177
Look for the project on social media (Twitter/X, LinkedIn) - profile images are often logos
- Distinctive: line 178
Search "[project name] logo" on image search engines
- Distinctive: line 179
Check DeFiLlama, CoinGecko, or similar aggregators for crypto projects
- Distinctive: line 180
Ask the user if they have alternative sources or contacts
- Distinctive: line 184
If the logo is rendered via JavaScript:
- Distinctive: line 186
Use MCP Playwright to load the page fully
- Distinctive: line 187
Wait for dynamic content to render
- Distinctive: line 192
If brand assets are behind a login:
- Distinctive: line 196
Ask the user to download manually or provide credentials (if appropriate)
≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Extract project logo
This skill extracts a project's official logo from its website or related sources, prioritising high-quality vector formats and official brand assets.
Required inputs
Before starting, gather the following from the user:
- Website URL - The project's main website (e.g.,
https://example.com) - Save path - Local filesystem path where the logo should be saved (e.g.,
./logos/example-logo.svg) - Format preference (optional) - Preferred format: SVG (recommended), PNG, or any
If any required input is missing, ask the user before proceeding.
Save all project logos in the target folder. We will pick variants and post-process those later.
There is no universal standard how artist name their logo files for dark and light variants. In our case, we always say
light: light (white) text on dark or transparent backgrounddark: dark (black) text on white or transparent background
Following vocabulary is used:
- Brand mark: same as logo mark, the logo without the brand name text
- Word mark: the logo with the brand name text
Step 1: Search for brand kit or media resources
The highest quality logos are typically found in official brand/media kits. Search for these pages:
- Navigate to the website and look for links in the footer or header:
- "Brand", "Brand Kit", "Brand Assets", "Brand Guidelines"
- "Media", "Media Kit", "Press", "Press Kit"
- "Resources", "Downloads", "Assets"
- Try common URL patterns:
{base_url}/brand{base_url}/brand-kit{base_url}/media{base_url}/press{base_url}/assets{base_url}/resources
- If a brand kit is found:
- Look for downloadable logo packages (often ZIP files)
- Prefer SVG or vector formats over raster images
- Choose the most square logo variant if multiple options exist
- If both light and dark themes are present, get both
- If colourful option exist, get it as well
If a brand kit is found with suitable logos, proceed to Step 5.
Step 2: Check GitHub repository
Many projects host their logos in their GitHub repositories:
- Find the project's GitHub repository (often linked in website footer/header)
- Search these common locations:
/assets/or/images/directories/branding/or/brand/directories/public/directory (for web apps)/docs/or/documentation/directories- Root directory (README badges, logo files)
- Check the README.md for embedded logos:
- Look for
<img>tags or markdown image syntax - These often point to high-quality logo files
- Look for
- Search for files named:
logo.svg,logo.png,logo-*.svg{project-name}.svg,{project-name}-logo.*brand.*,icon.*
If a suitable logo is found, proceed to Step 4.
Step 3: Extract from website directly
If no brand kit or GitHub assets are available, extract the logo from the website:
Option A: Using MCP Playwright (recommended for dynamic sites)
Use the MCP Playwright tool to:
- Navigate to the homepage
- Look for logo elements in the header (typically top-left):
- Search for
<img>tags with class/id containing "logo" - Check for
<svg>elements in the header - Look for elements with
role="img"and logo-related aria labels
- Search for
- Extract the image source URL or SVG content
Option B: Check meta tags and favicon
- Open Graph image - Check for
<meta property="og:image">tag- Often a high-quality image suitable for social sharing
- May include branding elements
- Favicon - Check for high-resolution favicon:
<link rel="icon" type="image/svg+xml">(best - vector)<link rel="apple-touch-icon">(usually 180x180 PNG)/favicon.ico(low resolution, last resort)
- Twitter Card image - Check
<meta name="twitter:image">
Step 4: Extract from Twitter
Get the project logo from their Twitter (also known as X.com) avatar image.
Step 5: Extract from Coingecko
Get the project logo from their Coingecko.
Step 6: Download and save
- For SVG files:
- Download the raw SVG content
- Ensure the file is valid XML
- Save with
.svgextension
- For raster images (PNG, etc.):
- Download the highest resolution available
- Verify file integrity after download
- Save with appropriate extension
- Verify the saved file:
- Open the file to confirm it displays correctly
- Check file size is reasonable (not 0 bytes or corrupted)
Download methods
Using curl:
curl -o "{save_path}" "{logo_url}"Using Python:
import requests
response = requests.get("{logo_url}")
with open("{save_path}", "wb") as f:
f.write(response.content)For SVG content extracted directly:
with open("{save_path}", "w") as f:
f.write(svg_content)Step 7: Report results to user
Provide the user with:
- Source location - Where the logo was found (brand kit URL, GitHub path, etc.)
- File details - Format, dimensions (for raster), file size
- Saved path - Confirm where the file was saved
- Alternative versions - Note if other variants are available (dark mode, icon-only, etc.)
Step 8: Save the report as Markdown file next to the logos
- Create a Markdown file in the folder where logos where saved
- Include the same information as for the user report
- Include links to the web pages, brand pages and brand kits used
Troubleshooting
No logo found
If no suitable logo can be found:
- Check if the project uses a different primary domain
- Look for the project on social media (Twitter/X, LinkedIn) - profile images are often logos
- Search "[project name] logo" on image search engines
- Check DeFiLlama, CoinGecko, or similar aggregators for crypto projects
- Ask the user if they have alternative sources or contacts
Dynamic/JavaScript-rendered logos
If the logo is rendered via JavaScript:
- Use MCP Playwright to load the page fully
- Wait for dynamic content to render
- Extract from the rendered DOM
Logo requires authentication
If brand assets are behind a login:
- Inform the user
- Provide the URL to the brand kit page
- Ask the user to download manually or provide credentials (if appropriate)
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

