evals playbooks for Claude Code, Codex and Cursor
2 reviewed playbooks of agent skills and MCP servers tagged evals: LangChain, LangGraph and LangSmith, Google ADK agents.
- LangChain, LangGraph and LangSmith: LangChain's own skills for building Deep Agents and LangGraph agents, tracing and monitoring them in LangSmith, and writing evals.
- Google ADK agents: Google's agents-cli skills for the Agent Development Kit: scaffold a project, write agents and tools, run evals, add tracing and deploy with approval.