JSON Schema Design

Define precise data contracts with JSON Schema for validation, documentation, and code generation.

Schema Fundamentals

Basic Types

{
  "$schema": "https://json-schema.org/draft/2020-12/schema",
  "$id": "https://organvm.dev/schemas/repo.json",
  "title": "Repository",
  "description": "An ORGANVM repository entry",
  "type": "object",
  "required": ["name", "organ", "tier", "status"],
  "properties": {
    "name": {
      "type": "string",
      "pattern": "^[a-z][a-z0-9-]*$",
      "minLength": 2,
      "maxLength": 64
    },
    "organ": {
      "type": "string",
      "enum": ["I", "II", "III", "IV", "V", "VI", "VII", "META"]
    },
    "tier": {
      "type": "string",
      "enum": ["flagship", "standard", "infrastructure"]
    },
    "status": {
      "type": "string",
      "enum": ["LOCAL", "CANDIDATE", "PUBLIC_PROCESS", "GRADUATED", "ARCHIVED"]
    },
    "tags": {
      "type": "array",
      "items": { "type": "string" },
      "uniqueItems": true
    }
  },
  "additionalProperties": false
}

Numeric Constraints

{
  "priority": {
    "type": "integer",
    "minimum": 1,
    "maximum": 10
  },
  "score": {
    "type": "number",
    "exclusiveMinimum": 0,
    "maximum": 1.0
  }
}

Composition

$ref (Reuse)

{
  "$defs": {
    "organ": {
      "type": "string",
      "enum": ["I", "II", "III", "IV", "V", "VI", "VII", "META"]
    },
    "timestamp": {
      "type": "string",
      "format": "date-time"
    }
  },
  "properties": {
    "source_organ": { "$ref": "#/$defs/organ" },
    "target_organ": { "$ref": "#/$defs/organ" },
    "created_at": { "$ref": "#/$defs/timestamp" }
  }
}

allOf (Intersection / Extension)

{
  "allOf": [
    { "$ref": "#/$defs/base-entity" },
    {
      "properties": {
        "extra_field": { "type": "string" }
      }
    }
  ]
}

oneOf (Discriminated Union)

{
  "oneOf": [
    {
      "properties": {
        "type": { "const": "skill" },
        "category": { "type": "string" }
      },
      "required": ["type", "category"]
    },
    {
      "properties": {
        "type": { "const": "bundle" },
        "includes": { "type": "array", "items": { "type": "string" } }
      },
      "required": ["type", "includes"]
    }
  ],
  "discriminator": { "propertyName": "type" }
}

if/then/else (Conditional)

{
  "if": {
    "properties": { "tier": { "const": "flagship" } }
  },
  "then": {
    "required": ["ci_url", "docs_url"]
  }
}

Patterns for Common Needs

Extensible Enums

{
  "status": {
    "anyOf": [
      { "enum": ["active", "archived", "draft"] },
      { "type": "string", "pattern": "^x-" }
    ]
  }
}

Maps / Dictionaries

{
  "metadata": {
    "type": "object",
    "additionalProperties": { "type": "string" },
    "propertyNames": { "pattern": "^[a-z_]+$" }
  }
}

Nullable Fields

{
  "description": {
    "oneOf": [
      { "type": "string" },
      { "type": "null" }
    ]
  }
}

Validation in Python

import jsonschema
import json
from pathlib import Path

def validate_entry(data: dict, schema_path: str) -> list[str]:
    schema = json.loads(Path(schema_path).read_text())
    validator = jsonschema.Draft202012Validator(schema)
    errors = sorted(validator.iter_errors(data), key=lambda e: list(e.path))
    return [f"{'.'.join(str(p) for p in e.path)}: {e.message}" for e in errors]

Schema Evolution

Change	Safe?	Strategy
Add optional field	Yes	No version bump needed
Add required field	No	Major version, provide default
Remove field	No	Deprecate first, then remove
Widen type (string → string|number)	Yes	Backward compatible
Narrow type	No	Major version
Add enum value	Yes	Consumers should handle unknown
Remove enum value	No	Deprecate first

Anti-Patterns

• No additionalProperties: false — Typos in field names pass silently
• Overly permissive types — Use specific types and constraints
• Inline definitions everywhere — Extract to $defs for reuse
• No $id or $schema — Always specify schema version and identity
• Validating only on write — Validate on both read and write boundaries