AOS Dashboard plugin
A local dashboard for Claude Code: a live graph of sessions and subagents, tokens, cost and file conflicts.
- 104 ★ on the repository
- Last commit
Extended access
Why this level
- Hooks write tool names, working folders and edited-file paths to a local log, and the monitor listens on a port on your machine.
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add plugins/aos-dashboardDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- Hooks write tool names, working folders and edited-file paths to a local log, and the monitor listens on a port on your machine.
This is third-party code. Review the repository files before installing.
What it does
AOS Dashboard watches running Claude Code sessions and never starts or stops anything. The plugin collects events through hooks and a local monitor shows agents, tasks, token usage, cost, conflicts when the same files are edited and alerts. The view is available as a streaming graph in the browser at `http://127.0.0.1:4318`, as a terminal view `aos-dash`, as a status line and as an MCP server. A Guard view shows verification verdicts for sessions and flags integrity rejections. An auto-verify hook on Stop is opt-in. Exact cost for the main agent and subagents can come from OpenTelemetry metrics, and across several machines the panel reads data over existing SSH access without new ports. There are no dependencies, only Node.js 18.
Who it is for. Developers who run several Claude Code sessions and subagents and want to see token spend and edit conflicts.
Good fit when
- Several subagents edit one repository and you need to spot file conflicts.
- You want to know which session burns tokens and what it costs.
- Work runs on several machines and you want one screen.
- You prefer agent status in the terminal through `aos-dash` without a browser.
Not a fit when
- You need to steer agents from the panel: it is read-only.
- Logging file paths and tool names in the local hook log is unacceptable.
- Work does not happen in Claude Code: the plugin collects data through its hooks.
Example request
Show which subagents are running now, how many tokens each session spent and whether there are file conflicts.Limitations
The author calls the project personal and macOS-first. Windows needs full Git for Windows and GNU make, and tests run only on Linux and macOS. Hooks do not log prompt text, file contents or tool output, but they record event and tool names, working folders, edited-file and transcript paths and a shortened task description. The page stays empty until a session sends events. The author moved the dashboard to the Fenceline project and new features land there, while this repository stays usable. The goal orchestrator is a separate experimental project.
How to disable. Run `make uninstall` in the repository folder: it removes the plugin and the `aos-dash` link and stops the monitor on port 4318. `make stop` stops only the monitor.
FAQ
How do I install the dashboard?
Clone the repository and run `make install`: the build takes about four seconds and the command installs the plugin, the `aos-dash` link and starts the monitor. `make doctor` checks the setup.
Why is the page empty?
The monitor shows data only after a new Claude Code session sends events. Start a session after installing and the graph fills in.
How do I lower the overhead?
Run `aos-dash` in the terminal and set the Streams button to Lines: no particles are drawn and per-frame redraws stop.
Can the dashboard stop agents?
No, it is read-only. Approving and starting goals is done by the separate experimental `aos-orchestrator` with a server on port 4319.
What about the Fenceline project?
The author moved the dashboard there and added rules with warnings. Fenceline's `make install` removes this plugin first.
Discussion
No questions or reviews yet.
Related
PluginsA local real-time dashboard for Claude Code and Codex sessions: agent activity and tool calls
A dashboard for every Claude Code session on the machine: live agents, cost, history search and MCP tools for Claude.
Local ledger of AI agent work: tokens, estimated cost, steps, changed files and checks per task for Claude Code, Codex and OpenCode
A plugin for Claude Code and Codex: a shared task wall, project memory, reports and a dashboard that shows agents' work across sessions
ClawMetry reads the session files agents already write and shows the timeline, tool calls, tokens and cost in one dashboard.
A top-style terminal dashboard for Claude Code and Codex sessions: spend, tokens, context load, processes and tool calls.
README in short
The repository belongs to a family of the AOS-1 specification, the dashboard and an experimental orchestrator. Docs cover architecture, overhead measurements, multi-machine use and the list of collected fields.