Menu
Steganographic USDC payment embedding - hide payment instructions inside images and audio
Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.
Based on SOC occupation classification
| name | usdc-whisper |
| version | 1.0.0 |
| description | Steganographic USDC payment embedding - hide payment instructions inside images and audio |
| author | Wizbisy |
Hide USDC payment requests inside images, memes, and audio files. Other agents extract and execute payments invisibly.
The Problem: Payment links are ugly, get blocked, and scream "transaction happening here."
The Solution: Embed payment data into any media. Share a meme, get paid. Send a podcast, include an invoice. The payment is invisible until an agent decodes it.
Payment Data → Compress → Encrypt → Embed in LSB of pixels/samples
{
"v": 1,
"type": "usdc_payment",
"chain": "base-sepolia",
"recipient": "0x...",
"amount": "10.00",
"memo": "For the artwork",
"expiry": 1707350400,
"sig": "0x..."
}
# Clone the skill
git clone https://github.com/wizbisy/usdc-whisper
cd usdc-whisper
npm install
const whisper = require('usdc-whisper');
// Create payment request
const payment = {
recipient: '0x984b62D817881a9A49ca242D15f3ACaFEd1e9375',
amount: '5.00',
memo: 'Thanks for the meme!',
chain: 'base-sepolia'
};
// Embed into image
const outputPath = await whisper.embed({
inputImage: './cool-meme.png',
payment: payment,
outputPath: './cool-meme-with-payment.png',
encryptionKey: 'optional-shared-secret' // Optional AES encryption
});
console.log('Payment embedded:', outputPath);
const whisper = require('usdc-whisper');
// Extract hidden payment
const payment = await whisper.extract({
imagePath: './received-meme.png',
encryptionKey: 'optional-shared-secret' // Must match if encrypted
});
if (payment) {
console.log('Found payment request:');
console.log(` To: ${payment.recipient}`);
console.log(` Amount: ${payment.amount} USDC`);
console.log(` Memo: ${payment.memo}`);
// Execute payment
if (await whisper.confirmPayment(payment)) {
const txHash = await whisper.executePayment(payment);
console.log('Paid:', txHash);
}
}
# Embed payment into image
usdc-whisper embed \
--image meme.png \
--recipient 0x984b62D817881a9A49ca242D15f3ACaFEd1e9375 \
--amount 5.00 \
--memo "Nice work!" \
--output meme-payable.png
# Extract and show payment
usdc-whisper extract --image meme-payable.png
# Extract and auto-pay (dangerous - use with caution!)
usdc-whisper extract --image meme-payable.png --auto-pay
# Batch scan directory for payments
usdc-whisper scan --dir ./downloads --recursive
The skill also runs as a service for agent-to-agent use:
curl -X POST http://localhost:3040/embed \
-F "image=@meme.png" \
-F "recipient=0x984b62D817881a9A49ca242D15f3ACaFEd1e9375" \
-F "amount=5.00" \
-F "memo=Thanks!"
curl -X POST http://localhost:3040/extract \
-F "image=@received-meme.png"
curl http://localhost:3040/health
Each pixel in a PNG has RGBA values (0-255 each). The last bit of each value is nearly invisible to humans but can store data:
Original pixel: R=10110110 G=11001010 B=01010101 A=11111111
Payment bit: 1 0 1 0
Modified pixel: R=10110111 G=11001010 B=01010101 A=11111110
The visual difference is imperceptible, but agents can read it perfectly.
Payment JSON is compressed with zlib before embedding:
Payments are signed by the requester:
const signature = await wallet.signMessage(
JSON.stringify({ recipient, amount, memo, expiry, nonce })
);
Extracting agent verifies signature matches before executing.
Add to your agent's workflow:
# In HEARTBEAT.md or agent workflow
- Check received images for hidden payments
- If found and valid, prompt human for approval
- Execute approved payments
// config.js
module.exports = {
chain: 'base-sepolia',
chainId: 84532,
rpc: 'https://sepolia.base.org',
usdc: '0x036CbD53842c5426634e7929541eC2318f3dCF7e',
explorer: 'https://sepolia.basescan.org'
};
MIT - Built for the USDC Hackathon by Wizbisy 🦞
"The best payment is one nobody sees coming."