Skip to content
| Marketplace
Sign in
Visual Studio Code>Visualization>Claude Code Agents MonitorNew to Visual Studio Code? Get it now.
Claude Code Agents Monitor

Claude Code Agents Monitor

edtroleis

|
1 install
| (0) | Free
List your Claude Code subagents and see which ones are running right now, from a dedicated sidebar view.
Installation
Launch VS Code Quick Open (Ctrl+P), paste the following command, and press enter.
Copied to clipboard
More Info

Claude Code Agents Monitor

CI Version Installs License: MIT

See your Claude Code subagents — and which ones are working right now — from a dedicated VS Code sidebar.

Claude Code lets you define subagents (specialized agents in ~/.claude/agents). This extension lists them and, when paired with a small hook, shows live execution state: which subagent is running, for how long, and what ran recently.

The Claude Agents view: running now, recent runs, and available agents

Features

  • Dedicated sidebar view with three sections:
    • Running now — active subagents with a live elapsed timer.
    • Recent — the last runs, marked completed or failed.
    • Available agents — every .md in ~/.claude/agents; click to open its definition.
  • Status bar indicator with the number of running agents.
  • Live updates — watches the status file and agents folder; no manual refresh needed.
  • Zero telemetry — everything is read from local files on your machine.

Requirements

  • VS Code 1.85.0 or newer.
  • Claude Code with subagents in ~/.claude/agents.
  • Node.js on your PATH (used by the optional status hook).

Getting started

  1. Install the extension.
  2. Open the Claude Agents view from the activity bar (robot icon). The Available agents section lists your subagents immediately.
  3. To populate Running now / Recent, enable the status hook below.

Enabling live execution (status hook)

The extension reads ~/.claude/agent-status.json. That file is produced by a Claude Code hook on the Task tool (the tool Claude uses to launch subagents). Add this to ~/.claude/settings.json:

{
  "hooks": {
    "PreToolUse": [
      {
        "matcher": "Task",
        "hooks": [
          { "type": "command", "command": "node ~/.vscode/extensions/edtroleis.claude-code-agents-monitor-*/hooks/agent-status-hook.js pre 2>/dev/null || true" }
        ]
      }
    ],
    "PostToolUse": [
      {
        "matcher": "Task",
        "hooks": [
          { "type": "command", "command": "node ~/.vscode/extensions/edtroleis.claude-code-agents-monitor-*/hooks/agent-status-hook.js post 2>/dev/null || true" }
        ]
      }
    ]
  }
}

The hook script is bundled with the extension. If you prefer, copy hooks/agent-status-hook.js anywhere stable and point the commands at that path instead. After editing settings, open the /hooks menu once (or restart Claude Code) so the config reloads.

The hook is deliberately defensive: it never fails the tool call and writes the status file atomically.

Settings

Setting Default Description
claudeAgentsMonitor.agentsDir ~/.claude/agents Folder scanned for subagent .md files.
claudeAgentsMonitor.statusFile ~/.claude/agent-status.json Status file written by the hook.

How it works

Claude Code (Task tool)
        │  PreToolUse / PostToolUse hook
        ▼
~/.claude/agent-status.json  ◄── written atomically by agent-status-hook.js
        │  fs.watch + light poll
        ▼
VS Code view + status bar

Privacy & security

  • No network calls, no telemetry. The extension only reads local files.
  • The hook only records the subagent type, a short description, the project folder name, and timestamps — no prompt contents or secrets.

See SECURITY.md for the threat model and how to report issues.

Development

npm install
npm run compile     # or: npm run watch
npm run lint
npm test
npm run package     # builds the .vsix

Press F5 in VS Code to launch the Extension Development Host.

Limitations

  • There is no official Claude Code API for "currently running agent"; state comes from the hook on the Task tool.
  • Parallel subagents of the same type in the same session may pair start/stop out of order (best-effort matching).

License

MIT © Edtroleis

  • Contact us
  • Jobs
  • Privacy
  • Manage cookies
  • Terms of use
  • Trademarks
  • Your Privacy Choices
  • Consumer Health Privacy
© 2026 Microsoft