Write and modernize Pydantic v2 models, validators, serialization, and settings in Python.
npx skills add https://github.com/bm629/agent-skills --skill pydantic-v2-bm629
Please help me install this Agent Skill. Skill: pydantic-v2 Source: https://github.com/bm629/agent-skills/tree/main/skills/pydantic-v2 Command: npx skills add https://github.com/bm629/agent-skills --skill pydantic-v2-bm629
class Config, .dict(), and @validator silently linger in codebases, producing deprecated or broken code. This Skill provides the current v2 API surface — ConfigDict, model_dump(), @field_validator, pydantic-settings — so generated or migrated code is correct and idiomatic. ## Core Features & Use Cases - Model definition and validation: Build BaseModel classes with typed fields, Annotated[..., Field(...)] constraints, @field_validator/@model_validator logic, and discriminated unions. - Serialization and configuration: Control output with model_dump* options, @field_serializer, @computed_field, aliases, and ConfigDict options like strict, frozen, and from_attributes. - Settings and migration: Load environment-based config via pydantic-settings, validate non-model shapes with TypeAdapter, and convert v1-era code using a full rename table. - Use Case: You inherit a service whose models still use class Config and .dict(). Ask the agent to modernize them — every model is rewritten to v2 idioms, with Optional defaults fixed and validators converted. ## Quick Start Use the pydantic-v2 skill to rewrite this v1-style model with class Config and .dict() calls into current Pydantic v2 idioms.High-intent search queries and answers about installing and using this skill.
Replace .dict() with model_dump(), .parse_obj() with model_validate(), @validator with @field_validator plus @classmethod, and the nested class Config with model_config = ConfigDict(...). Also fix Optional fields, which no longer get an implicit None default in v2.
Use @field_validator with @classmethod directly beneath it for single-field rules, and @model_validator(mode="after") operating on self for cross-field checks. Modes before, after, plain, and wrap control when validation runs relative to parsing.
No, BaseSettings moved to the separate pydantic-settings package in v2. Install it with pip install pydantic-settings and import BaseSettings and SettingsConfigDict from pydantic_settings to load configuration from environment variables and .env files.
In v2, Optional[int] only marks a field nullable; it does not supply a default, so the field remains required. Write int | None = None to make it both nullable and optional.
Use TypeAdapter to validate non-model shapes like list[Model], dicts, or primitives without defining a wrapper model. Build the adapter once and reuse it for hot paths, but never use it as a field annotation inside a BaseModel.
CodeAtCodeValidate and coerce Python data using type hints and Pydantic models.
manutejValidate customer support data with Pydantic V2 models and serialization.
PriyanshKuniyalAdd type hints, generics, and protocols to Python code for mypy and pyright.
alti3Standardize serialization and persistence plugin wiring in Litestar apps.