17Kernel Protocol

The underlying protocol for AI-native intent routing.

0. Protocol Vision

  • Vectorization: Transforming personal capabilities into high-dimensional vectors.
  • Privacy-First: Local encryption of raw data; only feature vectors are routed.
  • Intent Routing: Connecting nodes based on complementary skill sets, not just keywords.

1. Core Data Structure (MCP Node)

{
  "id": "uuid",
  "identity": {
    "name": "string",
    "background": "Full Stack Developer / Solopreneur",
    "tags": ["Vibe Coding", "System Architect"]
  },
  "vectors": {
    "logic": [0.12, 0.88, 0.45, ...], // Extracted from GitHub
    "aesthetic": [0.95, 0.22, 0.11, ...], // Extracted from Portfolio
    "intent": "Looking for: AI Hardware Engineer"
  },
  "mcp_data": {
    "super_power": "Builds MVP in 2 hours",
    "anti_power": "Hates writing CSS details"
  }
}

2. Matching Logic (Vibe Matcher)

The matching algorithm prioritizes complementary skill sets over similarity.

// Pseudo-code for Vibe Score Calculation

function calculateMatch(nodeA, nodeB) {
  // 1. Calculate Complementary Score (e.g., Dev + Designer)
  const skillComplement = crossProduct(nodeA.vectors.skills, nodeB.vectors.needs);
  
  // 2. Calculate Vibe Symbiosis (Personality fit)
  const vibeFit = cosineSimilarity(nodeA.vectors.vibe, nodeB.vectors.vibe);

  return (skillComplement * 0.7) + (vibeFit * 0.3);
}

3. Architecture

Feature Extractor

Analyzes unstructured data (GitHub, Twitter) using LLMs to generate standard 17Kernel vectors.

Intent Router

A distributed routing table that connects nodes based on current intent signals.