mlflow Claude Code Plugin
Claude Code Plugin Claude Code Monitoring AI Agents & AI App DevelopmentMonitoring & ObservabilityCode Quality, Review & TestingThe explanation below is AI-generated. Please verify it against the sources.
The mlflow plugin is a set of skills that let coding agents work directly with MLflow to instrument, trace, debug, evaluate, and monitor AI agent and LLM applications. It covers tasks such as adding tracing to application code, investigating failed traces, setting up evaluation datasets and scorers, querying usage metrics, and guiding new users through MLflow setup. According to the README, the skills are designed to work with multiple coding agents, including Claude Code, Cursor, Codex CLI, Gemini CLI, and OpenCode. An optional hook is also included that detects relevant prompts and suggests the matching skill automatically.
About the service
MLflow is an open-source platform for tracking, evaluating, and observing machine learning and generative AI applications. It provides a tracking server, experiment tracking, and storage for traces and metrics, and the skills in this plugin call into these MLflow capabilities to perform their tasks.
What you can do with mlflow
- Instrument Python or TypeScript code with MLflow tracing across frameworks such as OpenAI, Anthropic, LangChain, LangGraph, and LiteLLM
- Investigate a failed or problematic trace by examining spans and assessments
- Debug multi-turn chat sessions by reconstructing session history
- Set up and run an agent evaluation workflow with datasets and scorers, then review results
- Query aggregated metrics such as token usage, latency, and error rates with time-series breakdowns
- Get guided onboarding to MLflow based on your use case (GenAI apps or traditional ML)
- Search official MLflow documentation using the llms. txt index
Sources
History of mlflow
- Claude Code Plugin Added mlflow