Langfuse Skills
Official Langfuse skills: Claude Code and Cursor set up LLM tracing, manage prompts and datasets, and migrate projects from LangSmith
- 333 ★ on the repository
- Last commit
- MIT
Extended access
Why this level
- The Langfuse secret key opens reading and changing the project traces, prompts and datasets
- The migration skill moves live data between platforms
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add skills/langfuse-skillsDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- The Langfuse secret key opens reading and changing the project traces, prompts and datasets
- The migration skill moves live data between platforms
This is third-party code. Review the repository files before installing.
What it does
The Langfuse team ships two skills for its LLM observability platform. The main langfuse skill reads the langfuse.com documentation before writing code and uses `langfuse-cli` to reach data: it browses traces, prompts, datasets and scores. The `references/` folder holds separate guides: instrumenting an application, building a dataset, moving prompts into Langfuse, editing prompts, setting up evals, collecting user feedback, error analysis, judge calibration, cost and CI/CD. The second skill, migrate-to-langfuse, moves a project from another observability platform such as LangSmith, Arize or Braintrust. It first interviews the user about scope, then switches live tracing and transfers historical traces, datasets, prompts and evaluators. You need Langfuse project keys in the `LANGFUSE_PUBLIC_KEY`, `LANGFUSE_SECRET_KEY` and `LANGFUSE_BASE_URL` variables.
Who it is for. Teams that build LLM applications and keep traces, prompts and quality evals in Langfuse, cloud or self-hosted.
Good fit when
- An application in Python or TypeScript needs tracing for model calls
- Prompts live as strings in code and should move into Langfuse prompt management
- Traces exist but evals do not: you need a dataset and evaluators to start from
- The project runs on LangSmith or Braintrust and moves over with its trace history
- The release pipeline should stop when experiment results get worse
Not a fit when
- Observability lives in another platform and no move is planned
- You want general prompt advice without touching Langfuse data
- The project has no LLM calls, so there is nothing to instrument
- Company policy blocks the Langfuse API and you have no project keys
Example request
Add Langfuse tracing to our FastAPI service and move the prompts from code into prompt managementLimitations
- Langfuse needs an account, cloud or self-hosted, and three environment variables with project keys.
- The langfuse skill fetches documentation from langfuse.com, so access to that site must be open.
- The README documents install for Cursor and through `npx skills`; for other agents copy the folder by hand.
- The migrate-to-langfuse skill moves live data, so back up the source platform before it starts.
How to disable. Delete the langfuse and migrate-to-langfuse folders from the skills directory or disable the plugin in Cursor. Remove the Langfuse keys from the environment if you no longer need them.
FAQ
How do I install the Langfuse skills in Claude Code?
Run `npx skills add langfuse/skills --skill "langfuse"`. The migration skill installs separately if you drop the `--skill` flag and pick both.
Where do I get Langfuse keys?
In your Langfuse project open Settings, then API Keys. Put the public and secret key into environment variables together with the cloud or self-hosted address.
Does the skill work with a self-hosted Langfuse?
Yes, set your own address in `LANGFUSE_BASE_URL`. The README mentions the EU and US clouds and self-hosting separately.
Which platforms can migrate-to-langfuse move from?
The README names LangSmith, Arize and Braintrust and adds "etc.". The skill opens with an interview about the scope.
How does the skill beat reading the docs by hand?
It makes the agent load fresh langfuse.com pages first, since the SDK changes often, and then pick the fitting guide from `references/`.
Discussion
No questions or reviews yet.
Related
SkillsLangSmith skills
LangSmith Skills
Official LangChain skills for LangSmith: query and export traces, build eval datasets and write custom evaluators from your agent
A LangGraph 1.0 guide as a skill: ReAct agents, StateGraph, middleware, memory, MCP and RAG, so Claude Code writes LangChain code without outdated APIs
Official LangChain skills: choose between LangChain, LangGraph and Deep Agents, quickstarts, RAG, memory, human approval and eval engineering
A local MCP server that stores skills as versions and lets improved variants go live only after a human review
Prompt caching audit
LLM Cache Audit Skill
Skill for Codex and Claude Code that traces what breaks LLM prompt caching in code and config and proposes a verification step
A CLI and MCP server that turns doc sites, repos, PDFs and videos into skills and RAG-ready data
README in short
The README lists the two skills in a table, then the install options: Cursor plugin, the `skills` CLI and a manual symlink. A prerequisites section names three environment variables and where to get keys. It closes with example tasks and a link to discussions for requesting new skills.
README badge
Original SKILL.md
---
name: langfuse
description: >-
Interact with Langfuse and access its documentation: tracing, monitoring, creating datasets, running experiments, and evaluating AI applications. Use when needing to (1) query or modify Langfuse data, (2) look up Langfuse documentation, concepts, integration guides, a feature or SDK usage, or (3) do any AI engineering task (AI observability, prompt engineering/management, evaluation and evaluator management, experimentation, dataset management, evaluation-driven CI/CD, feedback collection). Invoke it for tasks in this scope even when Langfuse is not configured or explicitly mentioned.
allowed-tools:
- WebFetch(domain:langfuse.com)
- Bash(curl *langfuse.com/*)
- Bash(npx langfuse-cli api __schema *)
- Bash(npx langfuse-cli api * --help *)
- Bash(npx langfuse-cli api * list *)
- Bash(npx langfuse-cli api * get *)
- Bash(bunx langfuse-cli api __schema *)
- Bash(bunx langfuse-cli api * --help *)
- Bash(bunx langfuse-cli api * list *)
- Bash(bunx langfuse-cli api * get *)
---
# Langfuse
This skill helps you use Langfuse effectively across all common workflows: instrumenting applications, migrating prompts, debugging traces, and accessing data programmatically.
## Core Principles
Follow these principles for ALL Langfuse work:
1. **Documentation First**: NEVER implement based on memory. Always fetch current docs before writing code (Langfuse updates frequently) See the section below on how to access documentation.
2. **CLI for Data Access**: Use `langfuse-cli` when querying/modifying Langfuse data. See the section below on how to use the CLI.
3. **Best Practices by Use Case**: Read the relevant reference below use-case-specific guidelines before asking the user for more details or implementing.
4. **Use latest Langfuse versions**: Unless the user specified otherwise or there's a good reason, always use the latest version of Langfuse SDKs/APIs. Even if you're only creating a plan for another agent to execute, be explicit about the exact version to use.
5. **If you guide the user through UI** and are unsure about a label or location, inspect the user’s screenshots or ask to see the relevant screen. Do not assume UI labels have the exact same names as API, SDK, or CLI fields.
## Use case specific references
- instrumenting an existing function/application: references/instrumentation.md
- creating or getting to a good (evaluation) dataset to measure quality or test for regressions in AI systems: references/create-dataset.md
- migrating prompts from a codebase into Langfuse: references/prompt-migration.md
- creating a prompt or changing any part of an existing prompt, including small edits and debugging/tuning: references/prompt-engineering.md
- setting up evals when the user needs to identify gaps across signal capture, monitoring, and evaluator metrics ("I have traces, how do I set up evals?"): references/setting-up-evals.md
- capturing user feedback signals (explicit ratings, behavioral events, conversation signals, task outcomes) as scores: references/user-feedback.md
- further tips on using the Langfuse CLI: references/cli.md
- preparing a Langfuse project for the v4 platform migration: references/v4-project-migration.md
- calibrating a new or existing LLM-as-a-Judge against labeled examples, iterating on its prompt, and deploying the approved judge: references/judge-calibration.md
- systematic error analysis when requested directly or eval setup still lacks concrete failure modes after agent-led trace inspection: references/error-analysis.md
- setting up CI/CD experiment gates with `langfuse/experiment-action`: references/ci-cd.md
- optimizing the cost of an LLM application or AI agent: references/cost-optimization.md
- submitting feedback about this skill: references/skill-feedback.md
## 1. Langfuse API via CLI
Use the `langfuse-cli` to interact with the full Langfuse REST API from the command line. Run via npx (no install required):
Start by discovering the schema and available arguments:
```bash
# Discover all available resources
npx langfuse-cli api __schema
# List actions for a resource