mcpbeat

Langsmith Fetch

langchain-ai/langsmith-fetch

Fetches LangSmith traces for debugging agent behavior. Use when troubleshooting agent issues, reviewing conversation history, or investigating tool calls.

359 tokens
context cost
the whole folder, loaded on every use
1
files
instructions only
0
copies elsewhere
how many repositories repackaged it
2
stars on the repo
on the repository, not the skill itself

Install

one command, takes just this skill from the repository 
npx skills add https://github.com/langchain-ai/lca-skills --skill langsmith-fetch

The instruction itself

5 sections, as written by the author

Fetching LangSmith Traces

Requires langsmith-fetch in project dependencies and LANGSMITH_API_KEY in a .env file.

Setup

First, find the .env file containing LANGSMITH_API_KEY:

find . -name ".env" -type f 2>/dev/null | head -5

Commands

Use --env-file <path-to-.env> with all commands:

# Fetch recent traces (uses LANGSMITH_PROJECT from .env, or specify --project-uuid)
uv run --env-file <path> langsmith-fetch traces ./traces --limit 10
uv run --env-file <path> langsmith-fetch traces ./traces --project-uuid <uuid> --limit 10

# Fetch single trace by ID
uv run --env-file <path> langsmith-fetch trace <trace-id>

# Include metadata (timing, tokens, costs)
uv run --env-file <path> langsmith-fetch trace <trace-id> --include-metadata

Output Formats

  • --format pretty - Human-readable (default)
  • --format json - Pretty-printed JSON
  • --format raw - Compact JSON for piping

Troubleshooting Workflow

  • Find .env: find . -name ".env" -type f 2>/dev/null
  • Fetch recent traces: uv run --env-file <path> langsmith-fetch traces ./debug --limit 10
  • Find relevant trace in saved JSON files
  • Check: What tools were called? What did they return? Was it correct/expected?

How to use it

Copy the folder

Take langchain-ai/langsmith-fetch from the repository into ~/.claude/skills for personal use, or into .claude/skills inside a project.

Check the name does not clash

The agent identifies a skill by the name field in its header. Two skills with the same name cannot sit side by side — one of them will be ignored.