By tumblecat44
4-agent build team for Claude Code. Tell it what to build — PM writes the PRD, engineer implements, QA tests. Boundaries enforced by hooks, wave-based execution, research-backed design from 18 sources.
npx claudepluginhub tumblecat44/macrohard-teams --plugin macrohard-teamsEngineering Lead agent. Implements features in src/ based on PRD specs. Manages dependencies in package.json. Commits with Conventional Commits format.
Build orchestrator. Analyzes a feature request, decides if a team is needed, decomposes into wave-based tasks, delegates to product-pm/eng-lead/qa-tester, and verifies results. Use when the user wants to build a feature with a coordinated agent team.
Product Manager agent. Writes PRDs with user stories, acceptance criteria, and open questions. Writes only to docs/product/ and docs/specs/. Flags ambiguous requirements as questions instead of making assumptions.
QA Specialist agent. Writes tests in tests/ and e2e/ based on PRD acceptance criteria. Runs tests and reports results. Reports bugs but does not fix them.
Build a feature with a coordinated agent team. Takes a feature description and orchestrates PM, engineer, and QA through wave-based execution. Use when you want to build something non-trivial that needs PRD, implementation, and tests.
Show current build progress — what the team has built, which agents produced what, test results, and any pending work. Use to check what was done.
Claude Code plugin that gives you a 4-agent build team. Tell it what to build — PM writes the PRD, engineer implements, QA tests.
# 1. Add the marketplace
/plugin marketplace add tumblecat44/macrohard-teams
# 2. Install the plugin
/plugin install macrohard-teams@macrohard-teams
# 3. Activate (run once after install)
/reload-plugins
/macrohard-teams:build 유저 인증 시스템 만들어줘
The leader agent:
Check progress:
/macrohard-teams:status
| Agent | Model | Writes to | Enforced by |
|---|---|---|---|
| leader | Opus | orchestration only | never implements |
| product-pm | Sonnet | docs/product/, docs/specs/ | PreToolUse hook |
| eng-lead | Opus | src/, package* | PreToolUse hook |
| qa-tester | Sonnet | tests/, e2e/ | PreToolUse hook |
Boundaries are physically enforced — agents that try to write outside their directories get blocked with exit code 2.
Tasks run in dependency-based waves, not fixed stages:
| Hook | Event | Purpose |
|---|---|---|
| enforce-boundary.sh | PreToolUse (Write/Edit) | Blocks agents from writing outside their directories |
| enforce-commit.sh | PreToolUse (Bash) | Requires Conventional Commits format |
| log-decision.sh | PostToolUse (Write/Edit) | Logs every file change to decisions.jsonl |
| teammate-gate.sh | TeammateIdle | Blocks agents from going idle without deliverables |
| task-complete-gate.sh | TaskCompleted | Logs task completions |
| inject-context.sh | SubagentStart | Reminds agents of their boundaries on startup |
macrohard-teams/
├── .claude-plugin/
│ ├── marketplace.json # Marketplace catalog (source: ".")
│ └── plugin.json # Plugin manifest (v2.0.0)
├── agents/ # 4 agent specs
│ ├── leader.md
│ ├── product-pm.md
│ ├── eng-lead.md
│ └── qa-tester.md
├── hooks/
│ └── hooks.json # 6 hook events
├── scripts/ # 6 enforcement scripts
├── skills/
│ ├── build/SKILL.md # /macrohard-teams:build
│ └── status/SKILL.md # /macrohard-teams:status
└── README.md
Design decisions are backed by 18 research sources across 8 iterations with zero contradictions:
MIT
Complete collection of battle-tested Claude Code configs from an Anthropic hackathon winner - agents, skills, hooks, rules, and legacy command shims evolved over 10+ months of intensive daily use
Executes bash commands
Hook triggers when Bash tool is used
Modifies files
Hook triggers on file write and edit operations
Uses power tools
Uses Bash, Write, or Edit tools
Complete collection of battle-tested Claude Code configs agents, skills, hooks, rules, and legacy command shims evolved over 10+ months of intensive daily use
Complete collection of battle-tested Claude Code configs from an Anthropic hackathon winner - agents, skills, hooks, and rules evolved over 10+ months of intensive daily use
Complete collection of battle-tested Claude Code configs from an Anthropic hackathon winner - agents, skills, hooks, and rules evolved over 10+ months of intensive daily use
Comprehensive .NET development skills for modern C#, ASP.NET, MAUI, Blazor, Aspire, EF Core, Native AOT, testing, security, performance optimization, CI/CD, and cloud-native applications
Unity Development Toolkit - Expert agents for scripting/refactoring/optimization, script templates, and Agent Skills for Unity C# development