Pydantic Ai Testing

Test PydanticAI agents using TestModel, FunctionModel, VCR cassettes, and inline snapshots. Use when writing unit tests, mocking LLM responses, or recording...

This guide is coherent for its stated purpose, but before using it: 1) understand that running the recorded tests will call real LLM APIs and requires provider credentials in your environment — ensure keys are stored securely; 2) recorded cassettes and snapshots are files written to your repo and can contain prompts, responses, or other sensitive data — inspect git diff carefully before committing and scope recordings tightly; 3) prefer mocking (TestModel/FunctionModel) in CI or on sensitive projects rather than recording live API responses; 4) if a recording accidentally contains secrets, rotate the exposed credentials and remove the artifact from history. If you want me to, I can list the likely environment variables you’ll need for common providers (OpenAI, Anthropic, etc.) or suggest a safer local test workflow.

Type: OpenClaw Skill Name: pydantic-ai-testing Version: 1.0.1 The skill bundle is a documentation and tutorial resource for testing PydanticAI agents. It contains standard code snippets for mocking LLM responses, using VCR cassettes, and performing snapshot testing with legitimate libraries like pytest and inline-snapshot. The instructions in SKILL.md are aligned with the stated purpose of managing test artifacts and do not contain any malicious commands, data exfiltration attempts, or harmful prompt injections.