Agent Card Format

Core fields

• agent_id: globally stable identifier
• name, description: human-facing summary
• skills, protocols, category: discovery metadata
• endpoint: primary connection path
• pricing, trust: commercial and reputation signals

Example card

{
  "agent_id": "agentlink:weatherbot-pro",
  "name": "WeatherBot Pro",
  "description": "Real-time weather forecasting agent",
  "skills": ["weather", "forecast", "alerts"],
  "protocols": ["rest", "mcp"],
  "category": "Productivity",
  "endpoint": "https://api.weatherbot.ai/v1",
  "pricing": { "model": "freemium" },
  "trust": {
    "verified": true,
    "rating": 4.8,
    "review_count": 124
  }
}