# Pydantic

> Python data validation using type hints and runtime type checking with Pydantic v2's Rust-powered core.

- **Type:** Skill
- **Install:** `agentstack add skill-jartan-llc-grimoire-pydantic`
- **Verified:** Yes — security-reviewed for prompt injection and unsafe behavior
- **Seller:** [Jartan-LLC](https://agentstack.voostack.com/s/jartan-llc)
- **Installs:** 0
- **Category:** [AI & ML](https://agentstack.voostack.com/c/ai-and-ml)
- **Latest version:** 0.1.0
- **License:** MIT
- **Upstream author:** [Jartan-LLC](https://github.com/Jartan-LLC)
- **Source:** https://github.com/Jartan-LLC/grimoire/tree/main/plugins/pythonica/skills/pydantic

## Install

```sh
agentstack add skill-jartan-llc-grimoire-pydantic
```

Requires the [AgentStack CLI](https://agentstack.voostack.com/docs/cli). Works with Claude Code, Cursor, and any MCP-compatible agent.

## About

# Pydantic Validation Skill

## Quick Start

```python
from pydantic import BaseModel, Field, EmailStr
from datetime import datetime

class User(BaseModel):
    id: int
    name: str = Field(..., min_length=1, max_length=100)
    email: EmailStr
    created_at: datetime = Field(default_factory=datetime.now)
    is_active: bool = True

# Validate data
user = User(id=1, name="Alice", email="alice@example.com")
print(user.model_dump())  # {'id': 1, 'name': 'Alice', ...}

# Automatic type coercion
user2 = User(id="2", name="Bob", email="bob@example.com")
assert user2.id == 2  # String "2" coerced to int

# Validation error
try:
    User(id=3, name="", email="invalid")
except ValidationError as e:
    print(e.errors())
```

---

## Core Concepts

### BaseModel Foundation

```python
from pydantic import BaseModel, ConfigDict

class Product(BaseModel):
    model_config = ConfigDict(
        str_strip_whitespace=True,
        validate_assignment=True,
        use_enum_values=True,
        arbitrary_types_allowed=False
    )

    name: str
    price: float
    quantity: int = 0

# Usage
product = Product(name="  Widget  ", price=19.99)
assert product.name == "Widget"  # Whitespace stripped

# Validate on assignment
product.price = "29.99"  # Auto-converts to float
```

### Field Configuration

```python
from pydantic import Field, field_validator
from typing import Annotated

class Item(BaseModel):
    # Field constraints
    sku: str = Field(pattern=r'^[A-Z]{3}-\d{4}$')
    price: float = Field(gt=0, le=10000)
    stock: int = Field(ge=0, default=0)

    # Annotated types (Pydantic v2)
    quantity: Annotated[int, Field(ge=1, le=100)]

    # Descriptions and examples
    description: str = Field(
        ...,
        description="Product description",
        examples=["High-quality widget"]
    )

    # Deprecated fields
    old_field: str | None = Field(None, deprecated=True)

    @field_validator('sku')
    @classmethod
    def validate_sku(cls, v: str) -> str:
        if not v.startswith('ABC'):
            raise ValueError('SKU must start with ABC')
        return v
```

## Pydantic v2 Improvements

### Migration from v1

```python
# Pydantic v1
class OldModel(BaseModel):
    class Config:
        validate_assignment = True
        json_encoders = {datetime: lambda v: v.isoformat()}

# Pydantic v2
class NewModel(BaseModel):
    model_config = ConfigDict(
        validate_assignment=True,
        # json_encoders replaced by serializers
    )

    @model_serializer
    def ser_model(self) -> dict:
        return {...}

# Key changes:
# - .dict() -> .model_dump()
# - .json() -> .model_dump_json()
# - .parse_obj() -> .model_validate()
# - .parse_raw() -> .model_validate_json()
# - @validator -> @field_validator
# - @root_validator -> @model_validator
```

### Performance Improvements

```python
# v2 uses Rust core (pydantic-core) for 5-50x speedup
from pydantic import BaseModel
import time

class Data(BaseModel):
    values: list[int]
    names: list[str]

# Benchmark
data = {'values': list(range(10000)), 'names': ['item'] * 10000}
start = time.perf_counter()
for _ in range(1000):
    Data.model_validate(data)
elapsed = time.perf_counter() - start
print(f"Validated 1000 iterations in {elapsed:.2f}s")
```

## Field Types

### Built-in Types

```python
from pydantic import (
    BaseModel, EmailStr, HttpUrl, UUID4,
    FilePath, DirectoryPath, Json, SecretStr,
    PositiveInt, NegativeFloat, conint, constr
)
from typing import Literal
from pathlib import Path

class Example(BaseModel):
    # Email validation
    email: EmailStr

    # URL validation
    website: HttpUrl

    # UUID
    id: UUID4

    # File system paths
    config_file: FilePath
    data_dir: DirectoryPath

    # JSON string -> parsed object
    metadata: Json[dict[str, str]]

    # Secret (won't print in logs)
    api_key: SecretStr

    # Constrained types
    age: PositiveInt
    balance: NegativeFloat
    username: constr(min_length=3, max_length=20, pattern=r'^[a-z]+$')
    code: conint(ge=1000, le=9999)

    # Literal types
    status: Literal['pending', 'approved', 'rejected']
```

### Custom Types

```python
from pydantic import GetCoreSchemaHandler
from pydantic_core import core_schema
from typing import Any

class Color:
    def __init__(self, r: int, g: int, b: int):
        self.r, self.g, self.b = r, g, b

    @classmethod
    def __get_pydantic_core_schema__(
        cls, source_type: Any, handler: GetCoreSchemaHandler
    ) -> core_schema.CoreSchema:
        return core_schema.no_info_after_validator_function(
            cls.validate,
            core_schema.str_schema()
        )

    @classmethod
    def validate(cls, v: str) -> 'Color':
        if not v.startswith('#') or len(v) != 7:
            raise ValueError('Invalid hex color')
        r = int(v[1:3], 16)
        g = int(v[3:5], 16)
        b = int(v[5:7], 16)
        return cls(r, g, b)

class Design(BaseModel):
    primary_color: Color

# Usage
design = Design(primary_color='#FF5733')
assert design.primary_color.r == 255
```

## Validators

### Field Validators

```python
from pydantic import field_validator, model_validator

class Account(BaseModel):
    username: str
    password: str
    password_confirm: str

    @field_validator('username')
    @classmethod
    def username_alphanumeric(cls, v: str) -> str:
        if not v.isalnum():
            raise ValueError('must be alphanumeric')
        return v

    @field_validator('password')
    @classmethod
    def password_strong(cls, v: str) -> str:
        if len(v)  str:
        if not v or not v.strip():
            raise ValueError('must not be empty')
        return v.strip()
```

### Model Validators

```python
from pydantic import model_validator
from typing import Self

class DateRange(BaseModel):
    start_date: datetime
    end_date: datetime

    @model_validator(mode='after')
    def check_dates(self) -> Self:
        if self.end_date  dict:
        # Pre-processing before validation
        if isinstance(data, dict) and 'total' not in data:
            data['total'] = len(data.get('items', [])) * 10.0
        return data
```

### Root Validators (Wrap)

```python
from pydantic import model_validator, ValidationInfo

class Config(BaseModel):
    env: Literal['dev', 'prod']
    debug: bool = False

    @model_validator(mode='wrap')
    @classmethod
    def validate_config(cls, values: Any, handler, info: ValidationInfo):
        # Call default validation
        result = handler(values)

        # Post-validation logic
        if result.env == 'prod' and result.debug:
            raise ValueError('debug cannot be True in production')

        return result
```

## Type Coercion and Strict Mode

```python
from pydantic import BaseModel, ConfigDict, ValidationError

# Coercive mode (default)
class CoerciveModel(BaseModel):
    count: int
    price: float

data = CoerciveModel(count="42", price="19.99")
assert data.count == 42  # String -> int
assert data.price == 19.99  # String -> float

# Strict mode
class StrictModel(BaseModel):
    model_config = ConfigDict(strict=True)

    count: int
    price: float

try:
    StrictModel(count="42", price="19.99")  # Raises ValidationError
except ValidationError as e:
    print("Strict mode: no coercion allowed")

# Per-field strict mode
class MixedModel(BaseModel):
    flexible: int  # Allows coercion
    strict: Annotated[int, Field(strict=True)]  # No coercion

MixedModel(flexible="1", strict=2)  # OK
# MixedModel(flexible="1", strict="2")  # ValidationError
```

## Nested Models and Recursive Types

```python
from pydantic import BaseModel
from typing import ForwardRef

# Nested models
class Address(BaseModel):
    street: str
    city: str
    country: str

class Company(BaseModel):
    name: str
    address: Address

company = Company(
    name="ACME Corp",
    address={'street': '123 Main St', 'city': 'NYC', 'country': 'USA'}
)

# Recursive types (tree structure)
class TreeNode(BaseModel):
    value: int
    children: list['TreeNode'] = []

TreeNode.model_rebuild()  # Required for forward references

tree = TreeNode(
    value=1,
    children=[
        TreeNode(value=2, children=[TreeNode(value=4)]),
        TreeNode(value=3)
    ]
)

# Self-referencing with ForwardRef
class Category(BaseModel):
    name: str
    parent: 'Category | None' = None
    subcategories: list['Category'] = []

Category.model_rebuild()
```

## Generic Models

```python
from pydantic import BaseModel
from typing import Generic, TypeVar

T = TypeVar('T')

class Response(BaseModel, Generic[T]):
    success: bool
    data: T
    message: str = ''

class User(BaseModel):
    id: int
    name: str

# Usage with concrete type
user_response = Response[User](
    success=True,
    data=User(id=1, name='Alice')
)

# List response
list_response = Response[list[User]](
    success=True,
    data=[User(id=1, name='Alice'), User(id=2, name='Bob')]
)

# Generic repository pattern
class Repository(BaseModel, Generic[T]):
    items: list[T]

    def add(self, item: T) -> None:
        self.items.append(item)

user_repo = Repository[User](items=[])
user_repo.add(User(id=1, name='Alice'))
```

## Serialization

### Model Dump

```python
from pydantic import BaseModel, Field, field_serializer

class Article(BaseModel):
    title: str
    content: str
    tags: list[str]
    metadata: dict[str, Any] = {}

    # Serialization customization
    @field_serializer('tags')
    def serialize_tags(self, tags: list[str]) -> str:
        return ','.join(tags)

article = Article(
    title='Pydantic Guide',
    content='...',
    tags=['python', 'validation']
)

# Dump to dict
data = article.model_dump()
# {'title': 'Pydantic Guide', 'tags': 'python,validation', ...}

# Exclude fields
data = article.model_dump(exclude={'metadata'})

# Include only specific fields
data = article.model_dump(include={'title', 'tags'})

# Exclude unset fields
article2 = Article(title='Test', content='...', tags=[])
data = article2.model_dump(exclude_unset=True)  # metadata excluded

# By alias
class AliasModel(BaseModel):
    internal_name: str = Field(alias='externalName')

model = AliasModel(externalName='value')
model.model_dump(by_alias=True)  # {'externalName': 'value'}
```

### JSON Serialization

```python
from datetime import datetime
from pydantic import BaseModel, field_serializer

class Event(BaseModel):
    name: str
    timestamp: datetime

    @field_serializer('timestamp')
    def serialize_dt(self, dt: datetime) -> str:
        return dt.isoformat()

event = Event(name='Deploy', timestamp=datetime.now())

# Dump to JSON string
json_str = event.model_dump_json()
# '{"name":"Deploy","timestamp":"2025-11-30T..."}'

# Pretty print
json_str = event.model_dump_json(indent=2)

# Parse from JSON
event2 = Event.model_validate_json(json_str)
```

### Custom Serializers

```python
from pydantic import model_serializer

class User(BaseModel):
    id: int
    username: str
    password: SecretStr

    @model_serializer
    def ser_model(self) -> dict[str, Any]:
        return {
            'id': self.id,
            'username': self.username,
            # Never serialize password
        }

user = User(id=1, username='alice', password='secret123')
assert 'password' not in user.model_dump()
```

## Settings Management

### BaseSettings

```python
from pydantic_settings import BaseSettings, SettingsConfigDict
from pydantic import Field

class AppSettings(BaseSettings):
    model_config = SettingsConfigDict(
        env_file='.env',
        env_file_encoding='utf-8',
        env_prefix='APP_',
        case_sensitive=False
    )

    # Environment variables
    database_url: str
    redis_url: str = 'redis://localhost:6379'
    secret_key: SecretStr
    debug: bool = False

    # Nested settings
    class SMTPSettings(BaseModel):
        host: str
        port: int = 587
        username: str
        password: SecretStr

    smtp: SMTPSettings

# Reads from environment variables:
# APP_DATABASE_URL, APP_REDIS_URL, APP_SECRET_KEY, APP_DEBUG
# APP_SMTP__HOST, APP_SMTP__PORT, etc.

settings = AppSettings()
```

### Multi-Environment Settings

```python
from functools import lru_cache

class Settings(BaseSettings):
    environment: Literal['dev', 'staging', 'prod'] = 'dev'
    database_url: str
    api_key: SecretStr

    model_config = SettingsConfigDict(
        env_file='.env',
        extra='ignore'
    )

    @property
    def is_production(self) -> bool:
        return self.environment == 'prod'

@lru_cache
def get_settings() -> Settings:
    return Settings()

# Usage in FastAPI
from fastapi import Depends

@app.get('/config')
def get_config(settings: Settings = Depends(get_settings)):
    return {'env': settings.environment}
```

## FastAPI Integration

### Request/Response Models

```python
from fastapi import FastAPI, HTTPException
from pydantic import BaseModel, EmailStr

app = FastAPI()

class UserCreate(BaseModel):
    username: str = Field(min_length=3, max_length=50)
    email: EmailStr
    password: str = Field(min_length=8)

class UserResponse(BaseModel):
    id: int
    username: str
    email: EmailStr

    model_config = ConfigDict(from_attributes=True)

@app.post('/users', response_model=UserResponse)
def create_user(user: UserCreate):
    # FastAPI auto-validates request body
    # Returns only fields in UserResponse (password excluded)
    return UserResponse(
        id=1,
        username=user.username,
        email=user.email
    )
```

### Query Parameters

```python
from pydantic import BaseModel, Field
from fastapi import Query

class PaginationParams(BaseModel):
    skip: int = Field(0, ge=0)
    limit: int = Field(10, ge=1, le=100)

class SearchParams(BaseModel):
    q: str = Field(..., min_length=1)
    category: str | None = None
    sort_by: Literal['date', 'relevance'] = 'relevance'

@app.get('/search')
def search(params: SearchParams = Query()):
    return {'query': params.q, 'sort': params.sort_by}
```

### Response Model Customization

```python
class DetailedUser(BaseModel):
    id: int
    username: str
    email: EmailStr
    created_at: datetime
    last_login: datetime | None

@app.get('/users/{user_id}', response_model=DetailedUser)
def get_user(user_id: int, include_dates: bool = False):
    user = DetailedUser(
        id=user_id,
        username='alice',
        email='alice@example.com',
        created_at=datetime.now(),
        last_login=None
    )

    if not include_dates:
        return user.model_dump(exclude={'created_at', 'last_login'})
    return user
```

## SQLAlchemy Integration

### ORM Models with Pydantic

```python
from sqlalchemy import Column, Integer, String, DateTime
from sqlalchemy.orm import DeclarativeBase
from pydantic import BaseModel, ConfigDict

class Base(DeclarativeBase):
    pass

# SQLAlchemy ORM model
class UserDB(Base):
    __tablename__ = 'users'

    id = Column(Integer, primary_key=True)
    username = Column(String(50), unique=True)
    email = Column(String(100))
    created_at = Column(DateTime, default=datetime.utcnow)

# Pydantic model for validation
class UserSchema(BaseModel):
    model_config = ConfigDict(from_attributes=True)

    id: int
    username: str
    email: EmailStr
    created_at: datetime

# Usage
from sqlalchemy.orm import Session

def get_user(db: Session, user_id: int) -> UserSchema:
    user = db.query(UserDB).filter(UserDB.id == user_id).first()
    return UserSchema.model_validate(user)  # ORM -> Pydantic
```

### Hybrid Approach

```python
from pydantic import BaseModel

class UserBase(BaseModel):
    username: str
    email: EmailStr

class UserCreate(UserBase):
    password: str

class UserUpdate(BaseModel):
    username: str | None = None
    email: EmailStr | None = None
    password: str | None = None

class UserInDB(UserBase):
    model_config = ConfigDict(from_attributes=True)

    id: int
    created_at: datetime
    password_hash: str

# CRUD operations
def create_user(db: Session, user: UserCreate) -> Use

…

## Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

- **Author:** [Jartan-LLC](https://github.com/Jartan-LLC)
- **Source:** [Jartan-LLC/grimoire](https://github.com/Jartan-LLC/grimoire)
- **License:** MIT

Install and usage instructions live in the source repository linked above.

## Pricing

- **Free** — Free

## Security capabilities

Automated source analysis of v0.1.0 — what this tool can access:

- **Network access:** no
- **Filesystem access:** yes
- **Shell / process execution:** no
- **Environment & secrets:** yes
- **Dynamic code execution:** no

*"Yes" means the capability is present in the source — more access means more to trust, not that it is unsafe.*


## Versions

- **0.1.0** — security scan: passed — Imported from the upstream source.

## Links

- Listing page: https://agentstack.voostack.com/l/skill-jartan-llc-grimoire-pydantic
- Seller: https://agentstack.voostack.com/s/jartan-llc
- Browse the marketplace: https://agentstack.voostack.com/browse

---
Listed on AgentStack — the marketplace for AI agent skills and MCP servers. Every listing is security-reviewed. Creators keep 70%.