GitHub repositories · by kontrollable

I made a Claude Code mod that lets you see what your agents and subagents are doing live without leaving Claude.

Name: I made a Claude Code mod that lets you see what your agents and subagents are doing live without leaving Claude.
Author: kontrollable

A Reddit post introducing Apiarius, an open-source (MIT) Claude Code mod that adds a live "agent deck" showing every agent and subagent, their current activity, parent agent, inter-agent messages and changed files. It works via Claude Code hooks without altering Claude's behavior, and is available in the terminal, a local web page and the desktop app, with vector avatars for agents.

kontrollable@kontrollable

yusuflkq/apiarius

Translated

About this mod

Apiarius reads Claude Code's own events through hooks and displays a live deck of agents and subagents: what each is doing, who spawned it, messages between agents, and files each changed. It offers terminal, local web and desktop views, plus small vector avatars, and only intervenes when the user presses a prompt or send button. Free and open source under MIT at github.com/yusuflkq/apiarius; the author asks for feedback and help improving the terminal version.

Installation

Check the author's README for the marketplace and plugin name first. Commands may change as the repository evolves.

claude plugin marketplace add yusuflkq/apiarius
claude plugin install apiarius@apiarius

Original text / README

I mainly use Claude Code via the desktop app and with Opus 5.5 and the newer models delegating lots of work to subagents behind the scenes. Claude doesn't exactly make it easy to tell which ones are still working, which have finished and which are stuck going in circles.

So I built Apiarius, a Claude Code mod that adds a live agent deck that shows every agent and subagent, what each one is doing right now, who started it, what they tell each other and which files each one changed.

Then I brought it to the terminal (for the purists) and a local web page too, so you can watch from wherever you work.

Inspired a little by the latest agent apps like muse, grokbot and dots that give the agents a little personality and flair, so I gave them small vector avatars.

The single major principle I had to build this was that it shouldn’t change how you work or what Claude does. It just reads Claude Code’s own events through hooks and just shows you what’s happening. It only steps in with action if you press a prompt or a send button on the deck.

I built it with Claude Code itself, and watching the agents build their own deck via the deck within a couple hours which was trippy and made me realise how insane recursive self improvement will be.

In the video demo, Claude spins up a team of agents to ship v2 of a small mock API, then on a local web view they take a break and play a game of mafia/imposter/werewolf.

Free and open source (MIT): https://github.com/yusuflkq/apiarius

Let me know what else you want it to show, and if you main the terminal, give the terminal version a go and help me improve it.

Get a badge

Is this your project? Add this badge to your README or website. It links back to this page.

Size Small Large

Theme Light Dark

Markdown

[![Featured on ClaudeMods](https://claudemods.dev/badge/i-made-a-claude-code-mod-that-lets-you-see-what-your-agents--a4b834.svg)](https://claudemods.dev/en/builds/i-made-a-claude-code-mod-that-lets-you-see-what-your-agents--a4b834)

HTML

<a href="https://claudemods.dev/en/builds/i-made-a-claude-code-mod-that-lets-you-see-what-your-agents--a4b834"><img src="https://claudemods.dev/badge/i-made-a-claude-code-mod-that-lets-you-see-what-your-agents--a4b834.svg" alt="Featured on ClaudeMods" height="20" /></a>

Project information

Author: kontrollable
Use cases: Multi-agent orchestration
Added: Nov 7, 2026
Source updated: Nov 6, 2026
Stars: 13
Permission footprint: At least L2 (incomplete scan)
views: 17

dashboard hooks monitoring subagents terminal

Report this work Is this your work? Claim it

Permission footprint

Automatically scanned data. Review the source before installing. Levels describe capabilities, not a safety certification.

Incomplete scan.External dependency not scanned:; External dependency not scanned: fs The level shown is a lower bound; parts that could not be resolved may need higher permissions.

Read files
Write files
Intercept tool calls
Submit prompts
Execute commands
Environment variables

View matches (76）

Read files · line444 lineplugin/hooks/register.tsxconst stat = await $.fs.stat(dir, { resolve: true })

Write files · line688 lineplugin/hooks/register.tsxawait $.fs.write(`${await debugDirOf($)}/${name}.json`, `${JSON.stringify(state, null, 2)}\n`)

Intercept tool calls · line898 lineplugin/hooks/register.tsx

const ran = await $.tool.call({ tool: call.tool, ...call.input } as unknown as Parameters<EngineInterface['tool']['call']>[0])

Submit prompts · line1063 lineplugin/hooks/register.tsxvoid $.prompt.submit({ text: said, asUser: true }).then(

Execute commands · line1176 lineplugin/hooks/register.tsxconst first = /^.*?[.!?](?=\s|$)/.exec(said)?.[0] ?? said

Submit prompts · line1234 lineplugin/hooks/register.tsxvoid $.prompt.submit({ text: prompt, asUser: true }).then(

Read files · line1602 lineplugin/hooks/register.tsxconst stat = await $.fs.stat(path)

Read files · line1604 lineplugin/hooks/register.tsxreturn $.fs.read(path)

Read files · line1867 lineplugin/hooks/register.tsxkeep(await $.fs.read(path))

Read files · line1915 lineplugin/hooks/register.tsxconst text = await $.fs.read(`${root}/web/fixtures/state.v2.json`)

Environment variables · line1974 lineplugin/hooks/register.tsxoverride: await read(() => $.env.get('APIARIUS_SESSIONS')),

Environment variables · line1975 lineplugin/hooks/register.tsxapiariusHome: await read(() => $.env.get('APIARIUS_HOME')),

Environment variables · line1976 lineplugin/hooks/register.tsxconfigDir: await read(() => $.env.get('CLAUDE_CONFIG_DIR')),

Environment variables · line1977 lineplugin/hooks/register.tsxhome: (await read(() => $.env.get('USERPROFILE'))) ?? (await read(() => $.env.get('HOME'))),

Environment variables · line1978 lineplugin/hooks/register.tsxwtSession: await read(() => $.env.get('WT_SESSION')),

Environment variables · line1979 lineplugin/hooks/register.tsxtermProgram: await read(() => $.env.get('TERM_PROGRAM')),

Write files · line1983 lineplugin/hooks/register.tsxwrite: (path, text) => $.fs.write(path, text),

Read files · line1984 lineplugin/hooks/register.tsxlist: path => $.fs.list(path),

Read files · line1985 lineplugin/hooks/register.tsxstat: path => $.fs.stat(path),

Execute commands · line1988 lineplugin/hooks/register.tsxnode: args => $.process.run(['node', ...args], { timeoutMs: 10_000, ...(nodeCwd() ? { cwd: nodeCwd() } : {}) }),

Intercept tool calls · line2612 lineplugin/hooks/register.tsxon('tool.call', async ($, e, next) => {

Read files · line2764 lineplugin/hooks/register.tsx

if (plan.trim() === '' && typeof record?.filePath === 'string' && record.filePath !== '') plan = await $.fs.read(record.filePath)

Execute commands · line589 lineplugin/hooks/background.tsconst lead = CHDIR.exec(rest) ?? PS_ASSIGN.exec(rest) ?? ASSIGN.exec(rest) ?? JOIN.exec(rest)

Execute commands · line397 lineplugin/hooks/model.tsconst suffix = / \d+$/.exec(label)?.[0] ?? ''

Execute commands · line407 lineplugin/hooks/model.tsconst found = /^([^:]{1,18}?)\s*(?::\s+|\s[-–—]\s+)(\S.*)$/.exec(oneLine(description))

Execute commands · line1343 lineplugin/hooks/model.tsconst named = /permission to use (.+)$/i.exec(message.trim())?.[1]?.trim().toLowerCase()

Execute commands · line1460 lineplugin/hooks/model.tsconst said = new RegExp(`<${name}>([\\s\\S]*?)</${name}>`).exec(body)?.[1]?.trim()

Execute commands · line2119 lineplugin/hooks/model.tsfor (let found = stops.exec(line); found; found = stops.exec(line)) {

Execute commands · line2507 lineplugin/hooks/model.tsfor (let start = PASTE_START.exec(rest); start; start = PASTE_START.exec(rest)) {

Execute commands · line2511 lineplugin/hooks/model.tsconst next = ownAt >= 0 ? null : PASTE_END.exec(body)

Execute commands · line2574 lineplugin/hooks/model.tsconst name = ECHO_NAME.exec(text)?.[1]

Execute commands · line2575 lineplugin/hooks/model.tsreturn name ? commandOf({ name, args: ECHO_ARGS.exec(text)?.[1] ?? '' }) : ''

Execute commands · line3057 lineplugin/hooks/model.tsconst head = /^\s{0,3}(#{1,6})\s+(.*?)\s*#*\s*$/.exec(raw)

Execute commands · line3062 lineplugin/hooks/model.tsconst numbered = /^(\s*)\d+[.)]\s+(.*)$/.exec(raw)

Execute commands · line3067 lineplugin/hooks/model.tsconst bullet = /^(\s*)[-*+]\s+(?:\[[ xX]\]\s+)?(.*)$/.exec(raw)

Execute commands · line3174 lineplugin/hooks/model.tsconst found = /^[\w$.+-]+/.exec(s.slice(i, i + 256))?.[0]

Execute commands · line3248 lineplugin/hooks/model.tsconst found = /\bmeta\s*=\s*\{/.exec(script)

Execute commands · line3701 lineplugin/hooks/model.tsconst tool = isPermission ? /permission to use (.+)$/i.exec(needs.text.trim())?.[1]?.trim() : undefined

Execute commands · line3882 lineplugin/hooks/model.tsconst found = /<([a-z][\w-]*-message)\b([^>]*)>([\s\S]*?)<\/\1>/i.exec(text)

Execute commands · line3885 lineplugin/hooks/model.tsconst from = /\b(?:from|sender|teammate_id|teammate)\s*=\s*"([^"]*)"/i.exec(found[2] ?? '')?.[1]?.trim()

Execute commands · line280 lineplugin/hooks/web.tsconst family = /^([a-z]+)-(\d+)-(\d+)$/.exec(name)

Execute commands · line282 lineplugin/hooks/web.tsconst old = /^(\d+)-(\d+)-([a-z]+)$/.exec(name)

Execute commands · line284 lineplugin/hooks/web.tsconst major = /^(\d+)-([a-z]+)$/.exec(name)

Execute commands · line1797 lineplugin/hooks/web.tsconst at = /^@(\S+)\s*([\s\S]*)$/.exec(said)

Execute commands · line2025 lineplugin/hooks/web.tsreturn /^[pf]-([0-9a-f]{16})(?:-[0-9a-f]{8})?-[0-9a-f]{32}$/.exec(ticket)?.[1] ?? ''

Execute commands · line694 lineplugin/hooks/inspect.tsconst home = /^(?:[A-Za-z]:)?\/(?:Users|home)\/[^/]+(\/.*)$/i.exec(slashed)

Environment variables · line101 lineplugin/hooks/webServer.tsoverride: await read(() => $.env.get('APIARIUS_SESSIONS')),

Environment variables · line102 lineplugin/hooks/webServer.tsapiariusHome: await read(() => $.env.get('APIARIUS_HOME')),

Environment variables · line103 lineplugin/hooks/webServer.tsconfigDir: await read(() => $.env.get('CLAUDE_CONFIG_DIR')),

Environment variables · line104 lineplugin/hooks/webServer.tshome: (await read(() => $.env.get('USERPROFILE'))) ?? (await read(() => $.env.get('HOME'))),

Execute commands · line116 lineplugin/hooks/webServer.ts

const ran = await $.process.run(W.homeArgvOf(`${$.plugin.root.replace(/[\\/]+$/, '')}/web/home.js`, request), { timeoutMs: 10_000, ...(nodeCwd() ? { cwd: nodeCwd() } : {}) })

Write files · line359 lineplugin/hooks/webServer.tsawait $.fs.write(path, text)

Execute commands · line426 lineplugin/hooks/webServer.ts

const ran = await $.process.run(['node', '-p', 'process.ppid'], { timeoutMs: 10_000, ...(nodeCwd() ? { cwd: nodeCwd() } : {}) })

Execute commands · line515 lineplugin/hooks/webServer.tslet ran = await spawnWeb($, W.serverArgvOf(script, web.dir, web.pid, web.port, SR.registryDir()))

Execute commands · line518 lineplugin/hooks/webServer.tsran = await spawnWeb($, W.serverArgvOf(script, web.dir, web.pid, 0, SR.registryDir()))

Execute commands · line561 lineplugin/hooks/webServer.tsasync function spawnWeb($: EngineInterface, argv: readonly string[]): Promise<WebRun> {

Execute commands · line570 lineplugin/hooks/webServer.tsconst child = $.process.spawn({ argv, input: `${web.secret}\n`, ...(nodeCwd() ? { cwd: nodeCwd() } : {}) })

Intercept tool calls · line815 lineplugin/hooks/webServer.ts

const called = await $.tool.call({ tool: 'mcp__Claude_Browser__preview_start', url } as unknown as Parameters<EngineInterface['tool']['call']>[0])

Execute commands · line961 lineplugin/hooks/webServer.ts

const ran = await $.process.run(['node', `${$.plugin.root.replace(/[\\/]+$/, '')}/web/open.js`], { stdin: `${W.webUrlOf(web.port, ticket)}\n`, timeoutMs: 10_000, ...(nodeCwd() ? { cwd: nodeCwd() } : {

Read files · line124 lineplugin/hooks/gitRunner.tsreturn { place, stat: await $.fs.stat(joinPath(root, place)) }

Read files · line146 lineplugin/hooks/gitRunner.tstexts.push(await $.fs.read(one.abs))

Environment variables · line181 lineplugin/hooks/gitRunner.ts

if (userHome === null) userHome = ((await $.env.get('USERPROFILE').catch(() => undefined)) ?? (await $.env.get('HOME').catch(() => undefined)) ?? '').trim()

Execute commands · line228 lineplugin/hooks/gitRunner.tsconst ran = await $.process.run(argv, { env: G.GIT_ENV, timeoutMs: ms })

Read files · line294 lineplugin/hooks/gitRunner.tsnames = R.childDirsOf(await $.fs.list(dir))

Read files · line303 lineplugin/hooks/gitRunner.tsconst stat = await $.fs.stat(R.childOf(R.childOf(dir, name), '.git'))

Read files · line393 lineplugin/hooks/gitRunner.tsconst stat = await $.fs.stat(root, { resolve: true })

Read files · line422 lineplugin/hooks/gitRunner.tsconst stat = await $.fs.stat(abs, { resolve: true })

Execute commands · line463 lineplugin/hooks/gitRunner.ts

const ran = await $.process.run(FB.filesArgvOf(script), { stdin: FB.requestTextOf({ root, paths, most, textMost }), timeoutMs: 15_000, ...(nodeCwd() ? { cwd: nodeCwd() } : {}) })

Execute commands · line31 lineplugin/hooks/diff.tsconst header = HEADER.exec(line)

Execute commands · line183 lineplugin/hooks/diff.tsconst match = /^diff --git a\/(.*) b\/(.*)$/.exec(line)

Execute commands · line265 lineplugin/hooks/git.tsconst match = /^filter\.(.+)\.([^.]+)$/i.exec(key)

Execute commands · line457 lineplugin/hooks/git.tsconst match = /^(-|\d+)\t(-|\d+)\t([\s\S]*)$/.exec(part)

Execute commands · line514 lineplugin/hooks/git.tsconst files = /(\d+) files? changed/.exec(text)

Execute commands · line515 lineplugin/hooks/git.tsconst added = /(\d+) insertions?$\+$/.exec(text)

Execute commands · line516 lineplugin/hooks/git.tsconst removed = /(\d+) deletions?$-$/.exec(text)

Execute commands · line91 lineplugin/hooks/ui/v2/fit.tsconst share = typeof width === 'string' ? /^(\d+(?:\.\d+)?)%$/.exec(width) : null

I made a Claude Code mod that lets you see what your agents and subagents are doing live without leaving Claude.

About this mod

Installation

Get a badge

Similar projects

universal-modder: Making Mods for Single-Player Games with Claude Code

Recently made a Claude mod: for those used to developing with the matt skill, after calling implement-spec, I wanted to observe all agents’ progress, so I built a panel to view it

switchboard

next-steps-supervisor

I made a claude code mod that uses Haiku 5.5 (or Jev) to pick the best model and effort for your task. (and more...)

Claude Code with 82 skills can do backtesting and risk management; Claude Code can also become a co-pilot for quantitative trading. This toolset comes with 18 agen