PydanticAI and Pydantic in This Project¶
This section explains where schema-driven validation belongs and how it relates to the MCP server.
Scope Boundaries¶
- MCP server docs cover tool APIs, adapters, configuration, and runtime behavior in
src/solidworks_mcp/server.pyandsrc/solidworks_mcp/tools/. - Agent docs cover prompt workflows, specialist agents, and schema-validated outputs in
src/solidworks_mcp/agents/. - Planning docs cover future ideas and roadmap items in
docs/planning/.
Where Pydantic Is Used¶
- Tool input models: request validation for MCP tools (strict and typed inputs).
- Output contracts: predictable response payloads (
status,message, execution metadata). - Config models: runtime config and environment parsing.
Where PydanticAI Is Used¶
- Agent harness execution and model routing.
- Structured result schemas for agent responses.
- Recoverable failure handling with retry limits.
- Smoke tests for schema-conformant outputs.
Practical Rule¶
Use MCP docs when implementing or troubleshooting server tools. Use agent docs when designing prompt workflows, schemas, and logging/replay behavior. Use planning docs for proposals that are not shipped yet.