Should I use dataclasses or Pydantic models for dependencies?

Use dataclasses for simple resource containers and Pydantic models if you need complex validation or want to leverage Pydantic's ecosystem for your dependency structure.

Can I test agents without making real API or database calls?

Yes, this skill demonstrates the use of agent.override(deps=mock_deps) which allows you to swap real infrastructure for mocks during unit tests.

What is the main benefit of dependency injection in PydanticAI?

It decouples your agent logic from infrastructure, allowing tools and prompts to access external resources like databases and APIs in a type-safe, testable, and modular way.

How do I access dependencies inside an agent tool?

You include RunContext[DepsType] as an argument in your tool function. PydanticAI automatically passes the context, and you can access your resources via ctx.deps.

PydanticAI Dependency Injection

Name: PydanticAI Dependency Injection
Author: existential-birds

byexistential-birds

•

数据科学与机器学习

Implements structured dependency injection for PydanticAI agents to manage database connections, API clients, and external resources.

关于

This skill provides a standardized framework for implementing dependency injection in PydanticAI agents, enabling seamless access to external resources like database connections, API clients, and user contexts through RunContext. By leveraging type-safe dataclasses or Pydantic models, it ensures that your AI agents remain modular, testable, and robust. The skill covers the entire lifecycle of dependency management, from initial definition and runtime injection to advanced type-safe tool access and mocking dependencies for unit testing, making it an essential pattern for building production-grade AI services.

主要功能

14 GitHub stars
Type-safe dependency management using RunContext and deps_type
Full static analysis support with generic Agent type annotations
Support for dataclasses and Pydantic models for dependency definitions
Resource access within agent tools, system prompts, and dynamic instructions
Standardized patterns for mocking dependencies during testing environments

使用场景

Creating highly testable AI agents by swapping production infrastructure for mock objects
Implementing multi-tenant AI applications where user-specific context must be injected at runtime
Building agents that require persistent database connections or stateful API clients

关于

主要功能

14 GitHub stars
Type-safe dependency management using RunContext and deps_type
Full static analysis support with generic Agent type annotations
Support for dataclasses and Pydantic models for dependency definitions
Resource access within agent tools, system prompts, and dynamic instructions
Standardized patterns for mocking dependencies during testing environments

使用场景

Creating highly testable AI agents by swapping production infrastructure for mock objects
Implementing multi-tenant AI applications where user-specific context must be injected at runtime
Building agents that require persistent database connections or stateful API clients