sodiumsun

Agenttrail — AI skill for Claude Code

AI community

Local observability map for AI coding agents.

How to install Agenttrail

This entry records only its repository, not the path inside it, so there is no exact command to give. Open sodiumsun/agenttrail and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

What Agenttrail does

Local observability map for AI coding agents. Watch Claude Code, Codex, and Cursor plans, tool calls, file changes, and progress in real time.

Alternatives in AI

Manifest — Real-time cost observability for OpenClaw agents — track tokens, costs, messages, and model usage 4.1k ★
CodeIsland — Real-time AI coding agent status panel in your MacBook notch — live status, approvals & replies for 13 AI tool 2.3k ★
Memtrace Public — Structural memory for AI coding agents 466 ★

README

agenttrail

**Know what your coding agents are doing, while they are doing it.**

[![npm](https://img.shields.io/npm/v/agenttrail?color=e9a23b&label=npm)](https://www.npmjs.com/package/agenttrail) [![license](https://img.shields.io/badge/license-MIT-blue.svg)](LICENSE) [![node](https://img.shields.io/badge/node-%E2%89%A520-2dd4bf)](package.json) [![stars](https://img.shields.io/github/stars/sodiumsun/agenttrail?style=social)](https://github.com/sodiumsun/agenttrail)

Your coding agent has been working for half an hour. Is it making progress? Is it stuck? Did it quietly reopen the part it already called done?

agenttrail is a local, open-source observability layer for AI coding agents. It turns plans, tool calls, file changes, and progress from Claude Code, OpenAI Codex, Cursor, or any agent that edits files into a live project map.

Start an agent, walk away, and come back to this:

![agenttrail watching itself being built: a Claude session appears with its plan, edits stream live, and the demo task ticks green on camera](docs/demo.gif)

cd your-repo
npx agenttrail --open

That's it. No account, no global install, no telemetry. agenttrail opens on localhost and starts watching the repo.

How the live agent map works

A plan says what the agent intends to do. The filesystem says what it actually touched. agenttrail shows both.

Signal	What it tells you
Declared	The component and task the agent says it is working on
Observed	The files it is changing right now, including revisions to finished work

When those signals disagree, you know where to look. A completed card lights up when its files change again. A live run sits on the component it is touching. The map moves as the work moves.