GitHub repositories · by vedant-shah

clawd-hq

Name: clawd-hq
Author: vedant-shah

Turns your Claude Code chat into a chief of staff: keep talking while he hands work to specialists from your installed helpers and runs them in the background, with a live pixel-art office panel.

vedant-shah@vedant-shah

vedant-shah/clawd-hq/tree/main/plugins/clawd-hq

1 / 2

Translated

About this mod

When Claude is doing a long job in your main conversation, new requests pile up behind it, and changing an instruction means stopping the job and saying it again.

Clawd HQ turns your main chat into Clawd, a chief of staff. You brief him once. He picks the right specialist for each job from the helpers you already have installed (ECC, another kit, or your own agents), runs independent work in the background, and brings back one plain report. You keep talking the whole time, and you can change a running specialist's task midway without a restart.

Uses your own kit: a project gets a small team of the 5 to 10 helpers it actually needs.
Right-sized models: Haiku, Sonnet or Opus per job.
Pair-ups: two specialists check each other's work and come back with one conclusion.
Quiet chat: tool rows and hand-backs are folded away, so the conversation stays readable.
A live office panel shows who is working on what, with a progress bar for each specialist.

Permissions, in plain words: the plugin uses hooks to watch tool calls and draw the panel, reads past sessions on your machine to find which helpers you use, and asks Claude models a few small questions (each is listed in the README's Privacy section). Apart from those model calls, it opens no network connections and starts no processes. It writes to your project's CLAUDE.md only when you press "Add to CLAUDE.md".

Free, MIT licensed, unofficial. Needs Claude Code v2.1.287 or later. Type /office-demo for a 90-second tour.

Installation

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

claude plugin marketplace add vedant-shah/clawd-hq
claude plugin install clawd-hq@clawd-hq

Original text / README

When Claude is doing a long job in your main conversation, new requests pile up behind it, and changing an instruction means stopping the job and saying it again.

Uses your own kit: a project gets a small team of the 5 to 10 helpers it actually needs.
Right-sized models: Haiku, Sonnet or Opus per job.
Pair-ups: two specialists check each other's work and come back with one conclusion.
Quiet chat: tool rows and hand-backs are folded away, so the conversation stays readable.
A live office panel shows who is working on what, with a progress bar for each specialist.

Free, MIT licensed, unofficial. Needs Claude Code v2.1.287 or later. Type /office-demo for a 90-second tour.

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/clawd-hq-0299b6.svg)](https://claudemods.dev/en/builds/clawd-hq-0299b6)

HTML

<a href="https://claudemods.dev/en/builds/clawd-hq-0299b6"><img src="https://claudemods.dev/badge/clawd-hq-0299b6.svg" alt="Featured on ClaudeMods" height="20" /></a>

Project information

Author: vedant-shah
Use cases: Multi-agent orchestration
Added: Nov 8, 2026
Source updated: Nov 11, 2026
Stars: 6
Permission footprint: At least L3 (incomplete scan)
views: 28
Discoverers:: @@vedantshah2001
Discoverers:: @@vedantshah2001

background-agentschief-of-staffdelegationhooks orchestration pixel-art subagents workflow

Report this workClaimed by author

Permission footprint

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

Incomplete scan.External dependency not scanned: …-5-5; Dependency imports exceed the scan limit The level shown is a lower bound; parts that could not be resolved may need higher permissions.

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

View matches (46）

Execute commands · line123 lineplugins/clawd-hq/hooks/register.tsx

spawn: a => $.agent.spawn({ subagentType: SCOUT_TYPE, description: a.description, prompt: a.prompt }), list: () => $.agent.list(), version: async () => { const v = await $.session.version(); return v.

Intercept tool calls · line123 lineplugins/clawd-hq/hooks/register.tsx

spawn: a => $.agent.spawn({ subagentType: SCOUT_TYPE, description: a.description, prompt: a.prompt }), list: () => $.agent.list(), version: async () => { const v = await $.session.version(); return v.

Environment variables · line174 lineplugins/clawd-hq/hooks/register.tsx

const homeOf = async ($: EngineInterface) => ((await $.env.get("HOME").catch(() => undefined)) || (await $.env.get("USERPROFILE").catch(() => undefined)) || '').replace(/[\\/]+$/, '')

Read files · line177 lineplugins/clawd-hq/hooks/register.tsxroot: $.plugin.root, home, warn, read: path => $.fs.read(path), exists: path => $.fs.exists(path),

Write files · line213 lineplugins/clawd-hq/hooks/register.tsxwrite: (path, text) => $.fs.write(path, text), setSeed: (part, text) => $.store.set(SEED_KEYS[part], text),

Read files · line243 lineplugins/clawd-hq/hooks/register.tsx

hires: async () => hireTypes((await $.store.get(teamKey(cwd))) as SavedTeam | undefined), list: path => $.fs.list(path), fx: () => getFx($), setFx: fn => setFx($, fn),

Intercept tool calls · line253 lineplugins/clawd-hq/hooks/register.tsxcheck: i => $.tool.check(i as never), list: i => $.tool.call({ tool: 'Artifact', ...i }) as never,

Read files · line259 lineplugins/clawd-hq/hooks/register.tsx

read: path => $.fs.read(path), complete: ask => $.model.complete(ask), fillPrompt: text => knowIo($).fillPrompt(text), byOf: id => byOf($, id), get: key => $.store.get(key), set: (key, value) => $.sto

Model calls · line259 lineplugins/clawd-hq/hooks/register.tsx

read: path => $.fs.read(path), complete: ask => $.model.complete(ask), fillPrompt: text => knowIo($).fillPrompt(text), byOf: id => byOf($, id), get: key => $.store.get(key), set: (key, value) => $.sto

Read files · line268 lineplugins/clawd-hq/hooks/register.tsxexists: path => $.fs.exists(path), read: path => $.fs.read(path), write: (path, text) => $.fs.write(path, text),

Write files · line268 lineplugins/clawd-hq/hooks/register.tsxexists: path => $.fs.exists(path), read: path => $.fs.read(path), write: (path, text) => $.fs.write(path, text),

Model calls · line302 lineplugins/clawd-hq/hooks/register.tsxask: prompt => $.model.fork({ prompt }),

Model calls · line313 lineplugins/clawd-hq/hooks/register.tsxcomplete: ask => $.model.complete(ask),

Environment variables · line324 lineplugins/clawd-hq/hooks/register.tsx

const [now, home, config, known, p] = [await $.clock.now(), await homeOf($), await $.env.get('CLAUDE_CONFIG_DIR').catch(() => undefined), await read($, blurbs), await getPayroll($)]

Read files · line328 lineplugins/clawd-hq/hooks/register.tsx

now: () => now, cwd, home: () => home, isFolder: path => $.fs.stat(path).then(s => s.kind === 'dir', () => false), configDir: () => config || undefined, blurbs: () => known, allAgentTypes: () => types

Read files · line331 lineplugins/clawd-hq/hooks/register.tsx

list: dir => listEntries(dir, d => $.fs.list(d), path => $.fs.stat(path)), read: path => $.fs.read(path).catch(() => undefined), exists: path => $.fs.exists(path),

Model calls · line334 lineplugins/clawd-hq/hooks/register.tsxcomplete: async req => answerText(await $.model.complete(req)),

Read files · line347 lineplugins/clawd-hq/hooks/register.tsxfolder: () => cwd, root: () => $.session.root(), real: path => $.fs.stat(path, { resolve: true }).then(s => s.realPath),

Model calls · line451 lineplugins/clawd-hq/hooks/register.tsxask: prompt => $.model.fork({ prompt }), // the note's one short question, on the cached chat

Submit prompts · line459 lineplugins/clawd-hq/hooks/register.tsx

const councilIo = ($: EngineInterface): CouncilIo => ({ fill: text => $.prompt.fill({ text, mode: 'append' }), submit: text => $.prompt.submit({ text, asUser: true }), later: fn => $.clock.after(0, fn

Register tools · line516 lineplugins/clawd-hq/hooks/register.tsx['demo tool', () => $.tool.register(DEMO_SPEC)],

Register tools · line522 lineplugins/clawd-hq/hooks/register.tsx['offer rule tool', () => $.tool.register(OFFER_RULE_SPEC)],

Register tools · line524 lineplugins/clawd-hq/hooks/register.tsx['step tool', () => $.tool.register(STEP_SPEC)],

Register tools · line525 lineplugins/clawd-hq/hooks/register.tsx['find helpers tool', () => $.tool.register(FIND_HELPERS_SPEC)],

Register tools · line526 lineplugins/clawd-hq/hooks/register.tsx['contractor tool', () => $.tool.register(HIRE_CONTRACTOR_SPEC)],

Execute commands · line640 lineplugins/clawd-hq/hooks/register.tsx

if (spawnTypeOf(e) === SCOUT_TYPE) return handScoutSpawn(scoutHandoff, next.origin.plugin === 'clawd-hq', () => next(e)) as never // the setup scout: no brief notes, no desk

Intercept tool calls · line684 lineplugins/clawd-hq/hooks/register.tsxon('tool.call', { tool: 'Artifact' }, async ($, e, next) => {

Intercept tool calls · line692 lineplugins/clawd-hq/hooks/register.tsx

for (const tool of FILE_TOOLS) on('tool.call', { tool: tool as 'Edit' }, ($, e, next) => leftWatch(leftIo($), e as never, () => next(e))) // MultiEdit is no declared tool name, but the engine runs one

Intercept tool calls · line694 lineplugins/clawd-hq/hooks/register.tsxon('tool.call', async ($, e, next) => {

Execute commands · line67 lineplugins/clawd-hq/hooks/left-check.tsconst json = /\{[\s\S]*\}/.exec(reply)?.[0]

Execute commands · line144 lineplugins/clawd-hq/hooks/know.tsconst bold = /^\*\*(.+?)\*\*/.exec(kept)

Execute commands · line132 lineplugins/clawd-hq/hooks/left-scan.tsconst marker = comment === undefined ? undefined : MARKER.exec(comment)

Execute commands · line55 lineplugins/clawd-hq/hooks/panel-model.tsconst after = new RegExp(`${family}-(\\d{1,2})(?:-(\\d{1,2}))?(?!\\d)`, 'i').exec(id)

Execute commands · line56 lineplugins/clawd-hq/hooks/panel-model.tsconst before = new RegExp(`(\\d{1,2})(?:-(\\d{1,2}))?-${family}`, 'i').exec(id)

Execute commands · line64 lineplugins/clawd-hq/hooks/panel-model.tsconst family = FAMILY.exec(id)?.[1]

Execute commands · line133 lineplugins/clawd-hq/hooks/panel-model.tsconst end = CLAUSE_END.exec(rest)

Execute commands · line156 lineplugins/clawd-hq/hooks/panel-model.tsconst asked = WHEN_ASKED.exec(plain)

Execute commands · line38 lineplugins/clawd-hq/hooks/activity.tsconst host = /^https?:\/\/([^/?#:\s]+)/i.exec(text(input, 'url') ?? '')?.[1]

Execute commands · line57 lineplugins/clawd-hq/hooks/artifacts.tsreturn named ?? tidy(typeof text === 'string' ? /^\s*title:[ \t]*(.+)$/im.exec(text)?.[1] : undefined)

Execute commands · line135 lineplugins/clawd-hq/hooks/quiet.tsconst shortName = (tool: string) => MCP_NAME.exec(tool)?.[1] ?? tool

Execute commands · line59 lineplugins/clawd-hq/hooks/scout-run.tsconst started = await host.spawn({ subagentType: SCOUT_TYPE, description: 'Look around the project', prompt })

Execute commands · line96 lineplugins/clawd-hq/hooks/team.ts

const result = await io.spawnScout(brief).catch(() => ({ why: 'The scout could not start.' })) // the error's own words are not kept

Execute commands · line38 lineplugins/clawd-hq/hooks/usage-count.tsconst name = typeof content === 'string' && COMMAND_TEXT.test(content) ? COMMAND_NAME.exec(content)?.[1] : undefined

Execute commands · line78 lineplugins/clawd-hq/hooks/workflow-card.tsconst step = NUMBERED.exec(l)?.[1]

Execute commands · line90 lineplugins/clawd-hq/hooks/workflow-card.tsconst item = NUMBERED.exec(l)?.[1] ?? BULLET.exec(l)?.[1]

Execute commands · line53 lineplugins/clawd-hq/hooks/recap.tsconst at = NEXT_SENTENCE.exec(text)

clawd-hq

About this mod

Installation

Get a badge

Similar projects

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

agent-narrator

I built Project Sentinel : a Control Room Mod for long Claude Code runs

Microsoft quietly published 99 ready-made .NET skills for AI Agents

ruflo-agent