🜂 Manifest YAML

⚡ Quick Start

Create your first agent manifest in 5 minutes.

# manifest.yaml - Agent Identity Declaration
agent_id: my-agent-v1
name: "My First Agent"
version: "1.0.0"

identity:
  purpose: "Data analysis and reporting"
  creator: "user@example.com"
  timestamp: "2026-01-05T00:00:00Z"

permissions:
  read: ["public_data", "user_uploads"]
  write: ["reports"]
  execute: ["data_analysis"]

behavior_profile:
  mode: "supervised"
  autonomy_level: 0.6

trust:
  verification_required: true
  provenance_tracking: true
  mesh_coordination: true

# Install validator
npm install -g manifest-yaml-validator

# Validate manifest
manifest-yaml validate manifest.yaml

✓ Valid manifest (v1.0.0)
✓ Identity verified
✓ Permissions structured
✓ Trust substrate configured

# Register with manifest-yaml.com registry
curl -X POST https://manifest-yaml.com/api/registry \\
  -H "Content-Type: application/yaml" \\
  --data-binary @manifest.yaml

{
  "status": "registered",
  "agent_id": "my-agent-v1",
  "trust_score": 0.85,
  "registry_url": "https://manifest-yaml.com/agents/my-agent-v1"
}

What Is Manifest YAML?

The declarative layer that sits above agent identity cards and below transaction routing. Enables autonomous agents to declare not just who they are and what they can do — but under what terms they operate.

📚 Examples

agent_id: research-bot-v2
behavior_profile:
  mode: "autonomous"
  autonomy_level: 0.9
permissions:
  read: ["research_db", "papers"]
  execute: ["analysis", "citation"]

agent_id: assistant-v1
behavior_profile:
  mode: "supervised"
  autonomy_level: 0.3
permissions:
  read: ["calendar", "email"]
  write: ["notes"]

agent_id: swarm-node-42
behavior_profile:
  mode: "swarm-symmetric"
  autonomy_level: 0.7
trust:
  mesh_coordination: true
  swarm_id: "swarm-alpha"

📖 Specification

Manifest YAML v1.0 defines a standard format for agent identity declaration. Open specification, CC0 license, designed for agent-to-agent interoperability.

Core Components

• Agent Identity - ID, name, version, purpose, creator
• Permission System - Read, write, execute scopes
• Behavior Profile - Autonomy mode and level
• Trust Substrate - Verification and provenance
• Knowledge Substrate - Context and capabilities
• Mesh Coordination - Multi-agent communication

🔌 API & Registry

Public registry for agent manifest discovery and verification.

API Endpoints

# Register agent manifest
POST /api/registry
Content-Type: application/yaml

# Discover agent by ID
GET /api/agents/{agent_id}

# Search agents by capability
GET /api/search?capability=data-analysis

# Verify agent trust score
GET /api/verify/{agent_id}

Agent Declaration Layer v0.1

Policy, Authority, Jurisdiction, Evidence, and Audit for Agent-Readable Systems

A regulated agent cannot merely say "I can perform AML review." It must also declare: under what institutional authority, for which jurisdiction, using which data scopes, with what human-review requirements, with what audit log, and under what escalation rules. That declaration surface is currently unstandardized.

Declaration File Schema