ai-engineering-fluency: Skill for Claude Code

Refresh JSON Data Skill

This skill helps you update the token estimation ratios and model pricing data in the AI Engineering Fluency extension.

Overview

The extension uses two JSON data files that need periodic updates:

1. tokenEstimators.json - Character-to-token ratio estimators for AI models
2. modelPricing.json - Pricing information per million tokens for various AI models

These files are located in src/ directory and are bundled into the extension at build time.

When to Use This Skill

Use this skill when you need to:

• Add support for new AI models
• Update token estimation ratios based on new benchmarks
• Refresh pricing information from provider APIs
• Keep model data current with latest releases

Prerequisites

Before updating these files, ensure you have:

• Access to official pricing documentation from AI providers
• Token estimation benchmarks or documentation
• The repository cloned locally
• Node.js and pnpm installed

Step 1: Update tokenEstimators.json

Location

src/tokenEstimators.json

Structure

{
  "$schema": "http://json-schema.org/draft-07/schema#",
  "description": "Character-to-token ratio estimators for different AI models.",
  "estimators": {
    "model-name": 0.25
  }
}

Update Process

1. Research token ratios for new or updated models:

   • Look at the existing ratios for that provider's other models in tokenEstimators.json and match the pattern; verify with an actual tokenizer if unsure

2. Add or update entries in the estimators object:

   "new-model-name": 0.25
   

3. Validation:

   • Ensure JSON syntax is valid
   • Keep ratio values between 0.20 and 0.30
   • Use consistent formatting

Step 2: Update modelPricing.json

Location

src/modelPricing.json

Structure

{
  "$schema": "http://json-schema.org/draft-07/schema#",
  "description": "Model pricing data - costs per million tokens",
  "metadata": {
    "lastUpdated": "YYYY-MM-DD",
    "sources": [
      {
        "name": "Provider Name",
        "url": "https://pricing-url",
        "retrievedDate": "YYYY-MM-DD"
      }
    ],
    "disclaimer": "..."
  },
  "pricing": {
    "model-name": {
      "inputCostPerMillion": 1.25,
      "outputCostPerMillion": 10.0,
      "category": "Model category"
    }
  }
}