LangSmith skills
LangSmith Skills
Official LangChain skills for LangSmith: query and export traces, build eval datasets and write custom evaluators from your agent
Install
npx skills add langchain-ai/langsmith-skills --skill '*' --yesAdd --global to install for all projects.
This is third-party code. Review the repository files before installing.
What it does
Three skills help you observe and evaluate LLM apps. langsmith-trace adds tracing to code and exports traces for debugging through the langsmith CLI. langsmith-dataset builds evaluation datasets from traces. langsmith-evaluator creates custom evaluators. Each skill ships Python and TypeScript helper scripts.
Who it is for. For developers of LLM apps and agents who use LangSmith.
Good fit when
- You need to debug a wrong production answer from its traces
- You need a dataset for regression evals
- You need an evaluator for your own quality metric
Not a fit when
- Tracing lives in another observability tool
- You need skills for building with LangChain or LangGraph: use langchain-skills
Example request
Export yesterday's failing traces from the support-bot project and turn them into a datasetLimitations
Early development; content may change. Requires a LangSmith key, plus OpenAI or Anthropic keys for models. The trace skill needs the langsmith CLI installed.
How to disable. Uninstall the langsmith-skills plugin via /plugin or delete the skill folders from the agent's directory.
Security check
- Runs helper scripts and a CLI
- Reads traces that may contain user data
README in short
The README warns about early development and points to langchain-skills for agent building. It covers three install paths: skills.sh locally or globally, a Claude Code plugin and an install.sh script for Claude Code and Deep Agents CLI. Then it lists env vars and the three skills.
SKILL.md
--- name: langsmith-trace description: "INVOKE THIS SKILL when working with LangSmith tracing OR querying traces. Covers adding tracing to applications and querying/exporting trace data. Uses the langsmith CLI tool." --- <oneliner> Two main topics: **adding tracing** to your application, and **querying traces** for debugging and analysis. Python and Javascript implementations are both supported. </oneliner> <setup> Environment Variables ```bash LANGSMITH_API_KEY=lsv2_pt_your_api_key_here # REQUIRED LANGSMITH_PROJECT=your-project-name # Optional: default project LANGSMITH_WORKSPACE_ID=your-workspace-id # Optional: for org-scoped keys ``` Authentication is REQUIRED: either set the `LANGSMITH_API_KEY` environment variable, or pass the `--api-key` flag to CLI commands (preferred): ```bash langsmith trace list --project my-project --api-key $LANGSMITH_API_KEY ``` **IMPORTANT:** Always check the environment variables or `.env` file for `LANGSMITH_PROJECT` before querying or interacting with LangSmith. This tells you which project contains the relevant traces and data. If the LangSmith project is not available, use your best judgement to identify the right one. CLI Tool ```bash curl -sSL https://raw.githubusercontent.com/langchain-ai/langsmith-cli/main/scripts/install.sh | sh ``` </setup> <trace_langchain_oss>
FAQ
Which agents are supported?
Any agent supported by skills.sh, including Claude Code, Cursor, Windsurf and Goose. Claude Code also has a plugin.
How do I install globally?
Add --global to the npx skills add command.
Related
Open-source personal AI assistant on your own machine: answers in Telegram, Slack, Discord and WhatsApp, extended with skills and plugins
A self-improving agent from Nous Research with a TUI, messaging gateway, cron jobs and skills it writes itself
An open source coding agent for the terminal and desktop with build and plan modes
Open prompt library with a Claude Code plugin, MCP server and CLI: search, fetch and improve prompts and skills from an agent