---
title: "langwatch"
description: "Records which repository and branch each coding-agent session worked in, and teaches the agent to read its own traces back from LangWatch."
canonical: https://agentpluginsdirectory.com/plugins/langwatch
last-updated: 2026-09-22
---

# langwatch
Records which repository and branch each coding-agent session worked in, and teaches the agent to read its own traces back from LangWatch.
- Slug: langwatch
- Publisher: LangWatch
- Repository: https://github.com/langwatch/agent-plugin
- Manifest: plugin.json
- Version: 1.2.0
- License: MIT
- Category (editorial): agent-tooling
- Skills: 1 (langwatch)
- MCP servers: 0
- Stars: 0
- Repository created: 2026-08-08
- Repository last pushed: 2026-09-13
- Publisher type: Organization
- Listing: https://agentpluginsdirectory.com/plugins/langwatch
- Schema: https://agent-plugins.org/schemas/1.0.0/plugin.schema.json

## What langwatch does, in the publisher's words

One plugin directory, installable by two ecosystems.

- Claude Code reads.claude-plugin/plugin.json and the hooks beside it.
- Any Agent Plugins 1.0 client reads the root plugin.json and the skills/ directory.

Session context capture. At the start of a session and at the end of every turn, the plugin records which repository, branch and worktree the session is working in, and posts it to LangWatch as a single log record. That is what lets a coding-agent session's traces be joined to the code it worked on.

It reports repository identity only: origin remote host, owner and name, the current branch, and the worktree name. It does not read, send or index your code. It posts nothing when the CLI is not signed in, nothing outside a git repository, and nothing when the context has not changed since the last post. It never writes to the session's output and never fails a session.

From the project README, punctuation lightly normalized. Full text: https://raw.githubusercontent.com/langwatch/agent-plugin/HEAD/README.md

## Skills

- langwatch: Read LLM traces back from LangWatch with the langwatch CLI. Use when asked what an agent, prompt or model call actually did in production, when debugging a failed or slow LLM run, when looking up a trace or session by id, or when checking whether this coding session's own activity was captured.

Descriptions come from the frontmatter of each SKILL.md, punctuation lightly normalized.