langfuse — Claude Code plugin: tracing and prompts
Skills that let Claude query and manage Langfuse traces, prompts, datasets and scores through the API.
/plugin install langfuse@claude-plugins-official
/plugin marketplace add anthropics/claude-plugins-community
/plugin install langfuse@claude-community
/plugin marketplace add anthropics/claude-plugins-officialWhat langfuse does
The plugin brings two skills into Claude. The main one, langfuse, gives the agent access to the Langfuse API: it queries and manages traces, prompts, datasets and scores, looks up documentation, and tries to follow established practices while doing it. The second, migrate-to-langfuse, handles a move from another observability and evals platform such as LangSmith, Arize or Braintrust. It interviews you about scope first, then runs the live tracing cutover and transfers historical traces, datasets, prompts and evaluators.
You need a Langfuse account — cloud or self-hosted — and API keys, which live in your project under Settings > API Keys. Set them as environment variables: LANGFUSE_PUBLIC_KEY, LANGFUSE_SECRET_KEY and LANGFUSE_BASE_URL. Without those the agent cannot reach the API, so check them before anything else.
Skills load on their own when the task matches: setting up tracing in a project, auditing existing instrumentation, moving prompts into Langfuse prompt management, querying traces or datasets, looking up SDK and integration docs. There is also a path for CI/CD experiment gates via langfuse/experiment-action.
What it helps with
- set up Langfuse tracing in a project
- query and inspect traces through the API
- move prompts into Langfuse prompt management
- audit existing instrumentation for gaps
- migrate off LangSmith or Braintrust to Langfuse
Example prompts for the agent
- set up Langfuse tracing in this project
- pull the latest traces and show me where requests fail
- move our prompts into Langfuse prompt management
How to install langfuse
Plugin marketplace
/plugin install langfuse@claude-plugins-official/plugin marketplace add anthropics/claude-plugins-community/plugin install langfuse@claude-community/plugin marketplace add anthropics/claude-plugins-officialStep by step, with commands for each platform: Claude Code plugins: marketplaces, installing, and building your own
Questions about langfuse
How do I install langfuse in Claude?
From the marketplace: run /plugin marketplace add anthropics/claude-plugins-official, then /plugin install langfuse@claude-plugins-official.
What does the langfuse plugin do?
It adds two skills: one for working with the Langfuse API — traces, prompts, datasets, scores — and one for migrating from other observability platforms.
What do I need before using it?
A Langfuse account and API keys in the environment variables LANGFUSE_PUBLIC_KEY, LANGFUSE_SECRET_KEY and LANGFUSE_BASE_URL.
Source text
SKILL.md · MIT
<img width="2400" height="600" alt="hero-b" src="https://github.com/user-attachments/assets/c177a82d-334b-495c-81cb-9df69c521123" /> # Langfuse Skills [Agent Skills](https://github.com/anthropics/skills) that teach AI coding assistants (Claude Code, Cursor, etc.) how to work with [Langfuse](https://langfuse.com) — the open-source LLM engineering platform for tracing, prompt management, and evaluation. ## Skills | Skill | Description | | --------------------------------------------------------------- | -------------------------------------------------------------------------------------------------- | | [langfuse](./skills/langfuse) | Main skill to work with Langfuse. Query and manage traces, prompts, datasets, and scores via the Langfuse API; look up documentation; do things with best practices in mind. | | [migrate-to-langfuse](./skills/migrate-to-langfuse) | Migrate from another LLM observability/evals platform (LangSmith, Arize, Braintrust, etc.) to Langfuse. Interviews you about scope, then handles the live tracing cutover and transfer of historical traces, datasets, prompts, and evaluators. | ## Installation ### Cursor Plugin Install as a [Cursor plugin](https://cursor.com/docs/plugins): ``` /add-plugin langfuse ``` ### skills CLI Install via the [skills CLI](https://github.com/vercel-labs/skills): ```bash npx skills add langfuse/skills --skill "langfuse" ``` ### Manual symlink Clone this repo and symlink the skill into your agent's skills directory: ```bash git clone https://github.com/langfuse/skills.git /path/to/langfuse-skills ln -s /path/to/langfuse-skills/skills/langfuse /path/to/skills-directory/langfuse ``` ## Prerequisites You need a Langfuse account ([cloud](https://cloud.langfuse.com) or [self-hosted](https://langfuse.com/docs/deployment/self-host)) and API keys: ```bash export LANGFUSE_PUBLIC_KEY=pk-lf-... export LANGFUSE_SECRET_KEY=sk-lf-... export LANGFUSE_BASE_URL=https://cloud.langfuse.com # or https://us.cloud.langfuse.com, or self-hosted URL ``` API keys are found in your Langfuse project under **Settings > API Keys**. ## Usage Once installed, the agent will automatically use these skills when relevant — for example: - Setting up Langfuse tracing in a project - Setting up CI/CD experiment gates with `langfuse/experiment-action` - Auditing existing instrumentation - Preparing an application and Langfuse project for the v4 platform migration - Migrating prompts to Langfuse prompt management - Querying traces, prompts, or datasets via the API - Looking up Langfuse docs, SDK usage, or integration guides ## Contributing Want to improve a skill or add a new one? See [CONTRIBUTING.md](./CONTRIBUTING.md). ## Feedback & Requests Something not working as expected, or want a new skill? [Start a discussion](https://github.com/langfuse/skills/discussions/new?category=ideas-improvements).
© Langfuse · MIT · langfuse/skills
