Wrap ACP-compliant CLI agents as Pydantic AI model providers.
npx skills add https://github.com/arthrod/conejo-skills --skill pydantic-acp
Please help me install this Agent Skill. Skill: pydantic-acp Source: https://github.com/arthrod/conejo-skills/tree/main/skills/pydantic-acp Command: npx skills add https://github.com/arthrod/conejo-skills --skill pydantic-acp
This skill solves the integration gap between Pydantic AI and Agent Client Protocol (ACP) CLI tools, allowing you to use powerful agents like crush or cline as structured model backends while avoiding common pitfalls like JSON-RPC parsing errors and process leaks.
Use the pydantic-acp skill to initialize a new ACP model by providing the command list and a delegate client to handle host operations.
High-intent search queries and answers about installing and using this skill.
You can use ACP CLI agents as Pydantic AI models by wrapping them as model providers, which facilitates complex agentic tasks requiring structured output and tool execution via a delegate client.
The Agent Client Protocol (ACP) is used in Pydantic AI workflows to integrate ACP-compliant CLI agents, enabling host-level interaction and managing the session lifecycle and JSON-RPC communication.
JSON-RPC parsing errors and process leaks are common pitfalls when integrating ACP CLI tools; wrapping them as Pydantic AI models with a custom delegate client avoids these issues by managing the session lifecycle.
Yes, using Pydantic AI with ACP CLI agents requires the pydantic-acp package and a valid ACP stdio server implementation to manage the session lifecycle and JSON-RPC communication.
Yes, you can wrap specialized coding agents like crush-acp inside a Pydantic AI agent to perform complex, multi-step tasks that require both structured output and autonomous tool execution.
You handle filesystem, terminal, and permission operations through a custom delegate client, which provides a robust interface for managing host operations during the ACP agent session.
franklinbaldoIntegrate Pydantic AI ecosystem components for type-safe agents and workflows.
NikolasMarkouEnforce a state-machine-driven Explore, Plan, Execute, Reflect, and Pivot cycle for complex AI tasks.
EmeaAppGbbBuild and orchestrate AI agents with Microsoft Agent Framework.
EstomImplement MCP servers and clients with JSON-RPC 2.0 messaging and transports.
