Agent skill

Run Teamwork

by 0x0funky in 0x0funky/vibehq-hub

Run a single team session to build a project from a prompt. An agent skill from 0x0funky/vibehq-hub.

MITAuto-check passedAI & LLM Engineering

Install Run Teamwork

skills CLI
$ npx skills add 0x0funky/vibehq-hub --skill run-teamwork -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI
$ gh skill install 0x0funky/vibehq-hub run-teamwork --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Manual copy
$ git clone --depth 1 https://github.com/0x0funky/vibehq-hub.git skills-src && mkdir -p .claude/skills && cp -r skills-src/.claude/skills/run-teamwork .claude/skills/run-teamwork && rm -rf skills-src

Use ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.

Claude Code skills documentation · loads skills from .claude/skills/

Facts

Skill name
run-teamwork
GitHub stars
195
Token cost
~2.3k tokens
SKILL.md length
381 words
Files
1
Skills in repo
3
Repo updated
First seen
Licence
MIT

At a glance

Run a single team session to build a project from a prompt. An agent skill from 0x0funky/vibehq-hub.

  • Works in 4 steps: Parse & Design → Start → Wait for completion → …
  • Tasks that involve Prompt engineering
  • SKILL.md covers Step 1: Parse & Design, Step 2: Start, Step 3: Wait for completion and Step 4: Done, plus 1 more section
  • Calls node

What it does

Run Teamwork is an agent skill from 0x0funky/vibehq-hub. Run a single team session to build a project from a prompt. Designs the team, spawns agents, waits for completion, and delivers. No analysis or optimization loop.

Its SKILL.md is about 2.3k tokens, which your agent loads only when the skill is triggered. It is a single SKILL.md file with no bundled scripts.

It sits in AI & LLM Engineering, covering Prompt engineering. The repository describes itself as: Orchestrate Claude, Codex & Gemini agents working as a real engineering team. The licence is MIT.

When your agent uses it

  • Tasks that involve Prompt engineering

Example prompts

  • “/run-teamwork”

Workflow steps

4 steps, taken from the step headings in SKILL.md.

  1. Parse & Design
  2. Start
  3. Wait for completion
  4. Done

What it can do on your machine

Read from SKILL.md and the folder at commit 5f2964b. It shows what the files ask for, not the result of running them.

  • Tool permissions

    Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

    From allowed-tools in the SKILL.md frontmatter.

  • Runs code

    Shell commands in SKILL.md call:

    • node

    From the folder's file list and the shell code blocks in SKILL.md.

  • Network

    No URLs in SKILL.md.

    From URLs in SKILL.md, links to its own repository left out.

  • Credentials

    Names no API keys, tokens, secrets or passwords.

    From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Run Teamwork loads about 2.3k tokens when it runs. Until then it costs about 44 tokens; SKILL.md has 381 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~44
When it runs · the whole SKILL.md, loaded when a task matches
~2.3k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check passed

The automated check found no risky patterns in SKILL.md.

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from 0x0funky/vibehq-hub at commit 5f2964b, republished under its MIT licence (© 0x0funky). 381 words, ~2,258 tokens.

Download SKILL.mdSave it as .claude/skills/run-teamwork/SKILL.md (or your agent's skills folder).
name
run-teamwork
description
Run a single team session to build a project from a prompt. Designs the team, spawns agents, waits for completion, and delivers. No analysis or optimization loop.
argument-hint
"<project description>" [--port <number>]

/run-teamwork

You are a professional technical recruiter and team architect. Given a project description, you design the minimum effective team, spawn the agents, and let them build it. One shot — no analysis, no optimization loop.

Step 1: Parse & Design

Input: $ARGUMENTS contains the project description and optional flags.

  • Everything in quotes or before -- flags is the project prompt
  • --port <number> — hub port (default: 3013)
1a. Determine project name and team
  1. Project name — short kebab-case (e.g., chat-app, invest-tool)

  2. Count independent work domains — each domain = 1 agent = 1 directory

    Core principle: 1 agent = 1 independent work domain = 1 directory. Never put 2 agents in the same directory — they will overwrite each other's files.

    Decision framework:

    • Ask: "Would two agents ever need to edit the same file?" → If yes, merge into 1 agent
    • Ask: "Does this role have enough independent work to justify ~$10 cost?" → If no, merge
    • Prefer fewer, more capable agents over many specialized ones

    Sizing:

    DomainsTeamWhen
    1PM + 1API only, CLI, single-stack
    2PM + 2Typical full-stack (backend + frontend)
    3PM + 3Full-stack + separate data/infra domain
    4+PM + 4 maxLarge multi-stack. Cap at 5 total

    Anti-patterns:

    • ❌ 2 agents in same directory
    • ❌ Splitting one codebase by "feature" (shared types cause conflicts)
    • ❌ Designer agent without substantial UI deliverables
1b. Generate PM system prompt

Write to /tmp/vibehq-run-pm-prompt.md:

You are <Name>, the Project Manager for: <project prompt>

Project scope:
<break down into concrete deliverables>

Your workflow has TWO phases:

## Phase 1: Research
Before any implementation, create RESEARCH tasks for each domain that needs investigation.
Research tasks should ask team members to investigate and produce spec documents.

Examples of research tasks:
- "Research available free APIs for <domain>. Investigate endpoints, rate limits, auth requirements, response formats. Produce a spec document as a shared file: <domain>-research.md"
- "Research UX patterns and component libraries for <use case>. Produce ui-research.md"
- "Research best practices for <technical challenge>. Produce architecture-research.md"

Each research task MUST:
- Be assigned to the domain expert on the team
- Require a shared file as output (the spec/research document)
- Complete BEFORE any implementation tasks in that domain

## Phase 2: Implementation
After research tasks are done, READ the research output documents, then create implementation tasks.

Implementation tasks MUST:
- Reference the research output using `consumes` field
- Have specific acceptance criteria based on the research findings
- Require REAL integrations (real APIs, real libraries) — not mock/placeholder data
- Specify: "Mock data is only acceptable as a fallback when real API is unavailable"

## General rules:
1. Create a project brief first (publish_artifact)
2. Use depends_on to enforce: research tasks → implementation tasks
3. Use consumes to link implementation tasks to research output files
4. Track progress via list_tasks, unblock agents, ensure quality
5. When reviewing completed tasks: reject if using only mock data when real API was available
6. When all tasks are done, publish a final status report

Team:
<list each teammate with role and directory>

You are a COORDINATOR. Never write code. Only use MCP coordination tools.

For worker agents, do NOT generate custom system prompts — the spawner's built-in role presets are sufficient.

Show full SKILL.md (145 more words)Show less
1c. Print team design
========================================
Project: <project prompt>
========================================
Team: <project-name>
Port: <port>

Agents:
  - <Name> (PM) → <dir>/
  - <Name> (<Role>) → <dir>/<subdir>/
  ...

Spawning...
========================================

Step 2: Start

2a. Create directories
bash
mkdir -p <project-dir>/<subdir1> <project-dir>/<subdir2> ...

Project directory: ~/<project-name> (use home directory).

2b. Start hub
bash
node dist/bin/hub.js --port <port> --team <project-name> &

Run with run_in_background: true. Wait 3 seconds.

2c. Delete stale hub state
bash
rm -f ~/.vibehq/teams/<project-name>/hub-state.json
2d. Spawn agents

Write /tmp/vibehq-run-spawn.js:

javascript
const fs = require('fs');
const path = require('path');
const { exec, execSync } = require('child_process');

const config = <inline the team config as JSON>;
const isWindows = process.platform === 'win32';
const sessionName = `vibehq-${config.team}`;

function sleep(ms) { return new Promise(r => setTimeout(r, ms)); }

async function spawnAll() {
  if (!isWindows) {
    try { execSync(`tmux kill-session -t "${sessionName}" 2>/dev/null`); } catch {}
  }

  for (let i = 0; i < config.agents.length; i++) {
    const agent = config.agents[i];
    fs.mkdirSync(agent.cwd, { recursive: true });

    const args = [
      '--name', agent.name, '--role', agent.role,
      '--team', config.team, '--hub', `ws://localhost:${config.port}`,
      '--skip-permissions', '--auto-kickstart'
    ];
    if (agent.systemPromptFile) args.push('--system-prompt-file', agent.systemPromptFile);
    args.push('--', 'claude');

    console.log(`Spawning ${agent.name} (${agent.role})...`);

    if (isWindows) {
      const spawnCmd = `vibehq-spawn ${args.map(a => `"${a}"`).join(' ')}`;
      const launcher = path.join(require('os').tmpdir(), `vibehq-${agent.name.replace(/\s/g,'_')}-${Date.now()}.cmd`);
      fs.writeFileSync(launcher, `@echo off\r\nchcp 65001 >nul\r\nset CLAUDECODE=\r\ncd /d "${agent.cwd}"\r\n${spawnCmd}\r\npause\r\n`);
      exec(`wt -w new --title "${agent.name}" cmd /k "${launcher}"`);
    } else {
      const spawnCmd = `cd '${agent.cwd.replace(/'/g,"'\\''")}'&& CLAUDECODE= vibehq-spawn ${args.map(a=>`'${a}'`).join(' ')}; exec $SHELL`;
      if (i === 0) execSync(`tmux new-session -d -s "${sessionName}" -n "${agent.name}" "${spawnCmd.replace(/"/g,'\\"')}"`);
      else execSync(`tmux new-window -t "${sessionName}" -n "${agent.name}" "${spawnCmd.replace(/"/g,'\\"')}"`);
    }
    await sleep(3000);
  }

  if (!isWindows) {
    try {
      for (let w = config.agents.length - 1; w >= 1; w--)
        execSync(`tmux join-pane -s "${sessionName}:${w}" -t "${sessionName}:0" -h 2>/dev/null || true`);
      execSync(`tmux select-layout -t "${sessionName}:0" tiled`);
    } catch {}
    console.log(`\ntmux attach -t ${sessionName}`);
  }
}
spawnAll();

CRITICAL: Always include --auto-kickstart and CLAUDECODE= clearing.

Step 3: Wait for completion

Write /tmp/vibehq-run-poll.js:

javascript
const fs = require('fs');
const path = require('path');
const home = process.env.HOME || process.env.USERPROFILE;
const team = process.argv[2];
const statePath = path.join(home, '.vibehq', 'teams', team, 'hub-state.json');
if (!fs.existsSync(statePath)) { console.log('NO_STATE'); process.exit(0); }
const state = JSON.parse(fs.readFileSync(statePath, 'utf-8'));
const tasks = Object.values(state.tasks || {});
const total = tasks.length;
const done = tasks.filter(t => t.status === 'done' || t.status === 'rejected').length;
console.log(`Tasks: ${done}/${total}`);
for (const t of tasks) {
  const icon = t.status === 'done' ? 'v' : t.status === 'in_progress' ? '>' : t.status === 'rejected' ? 'x' : '.';
  console.log(`  [${icon}] ${t.title || t.id} -> ${t.status} (${t.assignee || '?'})`);
}
if (total > 0 && done === total) console.log('\nCOMPLETE');
else if (total === 0) console.log('\nNO_TASKS');
else console.log('\nWAITING');

Poll with sleep 30 && node /tmp/vibehq-run-poll.js <team> (60s timeout). Repeat until COMPLETE or 20 minutes elapsed.

Step 4: Done

When all tasks are complete, print a summary:

========================================
Project complete!
========================================
Team: <project-name>
Duration: <time>
Tasks: <done>/<total>
Output: <project-dir>/

View agents: tmux attach -t vibehq-<project-name>
Stop agents: tmux kill-session -t vibehq-<project-name>
========================================

Do NOT run analysis, optimization, or loop. The job is done.

Rules

  1. One shot — no iterations, no analysis, no loop.
  2. Always use --auto-kickstart — agents must start working immediately.
  3. tmux on macOS/Linux, Windows Terminal on Windows — never mix.
  4. Be patient — agents need 8-15 minutes typically.
  5. Hub must be started before agents — wait 3 seconds after hub start.
  6. Project directory in home dir — ~/<project-name>.

© 0x0funky, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

Just SKILL.md in .claude/skills/run-teamwork of 0x0funky/vibehq-hub.

Open the folder on GitHubat commit 5f2964b

Compare with similar skills

Run Teamwork next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

Run Teamwork compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Run Teamwork this skill0x0funky/vibehq-hub195—~2.3kAutomated safety check: PassMIT
Prompt Improverseverity1/claude-code-prompt-improver1.9k1 repos~1.7kAutomated safety check: PassMIT
Prompt Engineering Patternsynulihao/AgentSkillOS61714 repos~1.7kAutomated safety check: PassNone
Patch CreationPiebald-AI/tweakcc2.5k—~1.6kAutomated safety check: PassMIT
Senior Prompt Engineermaslennikov-ig/claude-code-orchestrator-kit2603 repos~1.4kAutomated safety check: PassCustom licence
Codex Fable5baskduf/FableCodex437—~1.6kAutomated safety check: PassAGPL-3.0

Similar skills

  • Prompt Improver

    severity1/claude-code-prompt-improver

    This skill enriches vague prompts with targeted research and clarification before execution.

    1.9k GitHub starsUsed in 1 repo~1.7k tokens
    AI & LLM EngineeringAuto-check passed
  • Prompt Engineering Patterns

    ynulihao/AgentSkillOS

    Master advanced prompt engineering techniques to maximize LLM performance, reliability, and controllability in production.

    617 GitHub starsUsed in 14 repos~1.7k tokens
    AI & LLM EngineeringAuto-check passed
  • Patch Creation

    Piebald-AI/tweakcc

    Create and register new patches for tweakcc. An agent skill from Piebald-AI/tweakcc.

    2.5k GitHub stars~1.6k tokensUpdated today
    AI & LLM EngineeringAuto-check passed
  • Senior Prompt Engineer

    maslennikov-ig/claude-code-orchestrator-kit

    Provides reference guides and Python scripts for prompt optimization, RAG evaluation, and agent orchestration when building or tuning LLM systems.

    260 GitHub starsUsed in 3 repos~1.4k tokens
    AI & LLM EngineeringAuto-check passed
  • Codex Fable5

    baskduf/FableCodex

    Apply a Claude Fable 5 inspired operating style inside Codex.

    437 GitHub stars~1.6k tokensUpdated 2 mo ago
    AI & LLM EngineeringAuto-check passed
  • Reference for designing and tuning production LLM prompts: few-shot examples, chain-of-thought, structured outputs, templates and system prompts.

    40k GitHub stars~1.3k tokensUpdated 4 days ago
    AI & LLM EngineeringAuto-check passed

More from 0x0funky/vibehq-hub

  • Optimize Protocol

    0x0funky/vibehq-hub

    Autonomous framework engineer — reads VibeHQ post-run analysis, understands root causes of multi-agent coordination failures, then designs and implements real code changes (new features, refactors…

    195 GitHub stars~3.2k tokensUpdated 6 mo ago
    Auto-check passed
  • Benchmark Loop

    0x0funky/vibehq-hub

    Fully automated self-improving loop — takes a project prompt, designs a team, runs the benchmark, analyzes results, optimizes framework code, rebuilds, and repeats until target grade is reached.

    195 GitHub stars~5.3k tokensUpdated 6 mo ago
    Auto-check: warnings

Questions about Run Teamwork

What does Run Teamwork do?

Run a single team session to build a project from a prompt. An agent skill from 0x0funky/vibehq-hub. Run Teamwork is an agent skill from 0x0funky/vibehq-hub. Run a single team session to build a project from a prompt.

When should I use Run Teamwork?

Run Teamwork fits situations like: tasks that involve Prompt engineering.

How do I install Run Teamwork in Claude Code?

Run `npx skills add 0x0funky/vibehq-hub --skill run-teamwork -a claude-code`. Or copy the skill folder (.claude/skills/run-teamwork in 0x0funky/vibehq-hub) into .claude/skills/run-teamwork in your project. Claude Code loads it when a task matches its description.

How do I install Run Teamwork in Codex?

Run `npx skills add 0x0funky/vibehq-hub --skill run-teamwork -a codex`. Or copy the skill folder (.claude/skills/run-teamwork in 0x0funky/vibehq-hub) into .agents/skills/run-teamwork in your project. Codex loads it when a task matches its description.

Can I use Run Teamwork in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add 0x0funky/vibehq-hub --skill run-teamwork -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/run-teamwork, .gemini/skills/run-teamwork, .github/skills/run-teamwork and .opencode/skills/run-teamwork in your project.

What does Run Teamwork need to run?

Going by SKILL.md and its folder, Run Teamwork needs the command-line tools its instructions call (node).

Does Run Teamwork access the network?

SKILL.md contains no URLs. Any network use would come from the scripts or tools the agent runs. This is read from the text; nothing was executed.

Is Run Teamwork safe to install?

Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.

What licence does Run Teamwork use?

Run Teamwork is published under the MIT licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Run Teamwork use?

About 2.3k tokens (SKILL.md is roughly 9k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.

What are the alternatives to Run Teamwork?

Skills that share tags, products or a category with Run Teamwork: Prompt Improver (severity1/claude-code-prompt-improver, 1.9k stars), Prompt Engineering Patterns (ynulihao/AgentSkillOS, 617 stars), Patch Creation (Piebald-AI/tweakcc, 2.5k stars) and Senior Prompt Engineer (maslennikov-ig/claude-code-orchestrator-kit, 260 stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Run Teamwork?

0x0funky (a GitHub user) maintains it in 0x0funky/vibehq-hub, which has 195 GitHub stars. The repository holds 3 skills in this directory. The repository was last updated on March 26, 2026.

Source: 0x0funky/vibehq-hub on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.