mlflowOfficial·11 Agent Skills Included

skills

Trace, debug, and evaluate LLM applications with MLflow

Instruments Python and TypeScript apps with MLflow tracing, then debugs failed traces and multi-turn chat sessions down to the exact span. Builds evaluation datasets and scorer suites, runs quality checks, and queries token usage, latency, and error metrics. Eliminates manual log digging and guesswork when fixing agent behavior, with regression tests that keep fixes from breaking later.

69|21

↗

npx skills add mlflow/skills --all -g -y

All Skills in This Repository (11)

Pure Emerald Level Indicators

📦 In Repo

mlflow

analyze-mlflow-trace

Analyze MLflow traces to inspect spans, attributes, and assessments.

analyze-mlflow-session

Analyze MLflow chat session traces to diagnose conversation issues.

mlflow-agent

Dispatches MLflow tracing, evaluation, and debugging workflows to appropriate sub-skills for AI agents.

searching-mlflow-docs

Search and retrieve official MLflow documentation via the llms.txt index.

sagemaker-mlflow

Connect SageMaker environments to MLflow for tracking experiments and managing models.

mlflow-onboarding

Guide users through MLflow setup and integration with Python code.

fix-agent-issue

Diagnose AI agent behavior issues by analyzing MLflow traces and creating regression tests.

querying-mlflow-metrics

Fetch and aggregate MLflow metrics with time-series and custom filters.

instrumenting-with-mlflow-tracing

Automate instrumentation of Python and TypeScript code for MLflow Tracing.

retrieving-mlflow-traces

Retrieve and search MLflow traces by ID, session, user, tags, and time.

agent-evaluation

Evaluate and optimize LLM agents using MLflow's evaluation API.

Official

Advanced

Frequently Asked Questions

FAQPage Schema

How to install MLflow Skills?▼

Run `npx skills add mlflow/skills --all -g -y` in your terminal to install all skills globally for your coding agent.

How to debug a failed LLM agent run?▼

Share the trace ID and the analyze-mlflow-trace skill walks every span, checks quality assessments, and pinpoints the exact step where things went wrong.

How to evaluate my AI agent's quality?▼

The agent-evaluation skill builds a test dataset, registers scorers, runs the full evaluation, and reports pass rates with concrete failure examples.

Does MLflow Skills work with Claude Code and Cursor?▼

Yes. All skills follow the universal SKILL.md standard and work in Claude Code, Cursor, Codex CLI, Gemini CLI, and OpenCode.

Can I track token usage and LLM costs?▼

Yes. The querying-mlflow-metrics skill fetches token usage, latency percentiles, and error rates with time-series breakdowns from your tracking server.