How do I track score regressions with this skill?

You can use the regression command to compare a historical baseline period against a current period to identify any statistically significant drops in model performance.

What credentials are required to use this skill?

You need to provide your Langfuse Public Key, Secret Key, and Host URL as environment variables to authenticate the analyzer.

What is Langfuse Score Analytics?

It is a Claude Code skill that enables automated analysis of LLM evaluation scores and quality metrics stored in the Langfuse observability platform.

What dimensions can I compare scores by?

The skill allows you to compare scores across different releases, environments (e.g., prod vs. dev), and trace names.

Can I analyze custom score names like 'helpfulness' or 'accuracy'?

Yes, you can specify any custom score name defined in your Langfuse project when running analysis commands.

Langfuse Score Analytics

Name: Langfuse Score Analytics
Author: mberto10

bymberto10

分析与监控

Analyzes and visualizes LLM quality scores, trends, and regressions within the Langfuse observability platform.

关于

Langfuse Score Analytics is a specialized Claude Code skill designed for developers and LLM engineers who need deep insights into their model performance. It provides automated tools to track score metrics over time, identify performance regressions between releases, and compare quality across different environments or trace names. By integrating directly with Langfuse, it allows users to quickly visualize distributions and trends, making it easier to monitor AI application reliability and optimize model outputs based on quantitative data directly from the CLI.

主要功能

0 GitHub stars
Analyze score trends with configurable time granularity
Compare quality metrics across releases, environments, or trace names
Detect performance regressions by comparing baseline and current windows
Visualize score distributions using histogram-style binning
Summarize and list all available evaluation scores in a Langfuse project

使用场景

Monitoring LLM performance shifts after deploying a new prompt or model version
Identifying quality discrepancies between staging and production environments
Quantifying distribution of user feedback scores to prioritize model improvements

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add mberto10/mberto-compound langfuse-score-analytics

For use in Claude.ai and ChatGPT

Download Skill

GitHub

关于