AI agents for Claude Code — team orchestration, scientific visualization, and local LLM delegation
npx claudepluginhub braininahat/brains-in-a-hat20 named specialist agents with Neal as chief of staff, agent teams coordination, vault persistence, and live dashboard
Production-ready workflow orchestration with 79 focused plugins, 184 specialized agents, and 150 skills - optimized for granular installation and minimal token usage
Claude Code marketplace entries for the plugin-safe Antigravity Awesome Skills library and its compatible editorial bundles.
Curated collection of 129 specialized Claude Code subagents organized into 10 focused categories
A 20-agent AI software team for Claude Code. Smart routing, auto-CODEOWNERS, self-improvement, and proactive DX suggestions.
# Install the team plugin
claude plugins add /path/to/brains-in-a-hat/brains-in-a-hat
# Install the visualization plugin (optional)
claude plugins add /path/to/brains-in-a-hat/grant
| Plugin | Purpose |
|---|---|
brains-in-a-hat | 20-agent dev team — routing, QA, memory, retrospectives |
grant | Visualization agent — architecture diagrams, git activity, code maps |
| Agent | Role |
|---|---|
| Session Manager | Briefings at session start, memory updates at end |
| QA Engineer | Advisory QA reviews before commits |
| Meta/Retro | Retrospectives, DX suggestions, CODEOWNERS maintenance |
| Agent | Trigger |
|---|---|
| System Designer | New features, architectural decisions |
| Architect | Code review for boundary violations |
| UI Reviewer | QML/CSS/HTML visual changes |
| Qt/QML | PySide6/Qt-specific patterns |
| MLOps | ONNX models, inference, weights |
| Signal Processing | Audio/video pipelines, timestamps |
| Hardware/Device | Probe protocol, WiFi, USB, V4L2 |
| Domain Expert | Clinical/domain-specific validation |
| Profiler | Performance, latency, memory |
| Researcher | Technical investigation, benchmarks |
| Data/Schema | SQLite, migrations, config |
| UX/Workflow | End-to-end user flows |
| Testing Strategy | Test suite design, coverage |
| Agent | Role |
|---|---|
| DevOps | CI/CD, GitHub Actions |
| Packaging | AppImage, Docker, PyInstaller |
| Docs Writer | Specs, CLAUDE.md, API docs |
| Parker | Issues, backlog, milestones, GitHub Projects |
/assemble — Activate the brains-in-a-hat team (Neal + all agents)/team-briefing — Session start briefing: branch status, uncommitted changes, open tasks/team-debrief — Session end memory update: save decisions, WIP, workflow changes/team-retro — Post-task retrospective: what went well, what to improve/team-review — QA review of staged/modified changes before commit/team-cleanup — Find dead code, unused imports, stale TODOs/grant — Visualize architecture, git activity, code structure, or agent topologyThe dashboard auto-starts via the session-start hook each time a Claude Code session opens. It requires Python 3 or uv to be available on $PATH.
http://localhost:8787 by default?project=<path> query parameter for multi-project use.brains_in_a_hat/state/dashboard.log in the current projectCopy examples/user-preferences.json to .brains_in_a_hat/user-preferences.json in your project and customize. Controls communication style, tool preferences, engineering principles. The Meta Agent also learns and updates this file over time.
Copy examples/domain-config.json to .brains_in_a_hat/domain-config.json and customize for your domain. The included example is for clinical speech-language pathology — replace with your own domain terminology, compliance requirements, and validation rules.
The plugin lives in a Git repo — install on any device with Claude Code. Project-specific config (CODEOWNERS, domain-config) stays in .brains_in_a_hat/. Vault artifacts persist globally at ~/.brains_in_a_hat/vault/.
MIT