By seongsu-kang
Orchestrate multi-agent AI teams (CTO, Engineer, QA, Designer) directly from Claude Code. Create custom agencies via a guided flow, install from a marketplace, and manage collaborative work waves with role-specific status tracking, task boards, cost breakdowns, and activity logs.
Create a custom agency with guided setup — scans project, suggests team, auto-generates, auto-verifies
Install an agency from marketplace or GitHub
List installed agencies
Publish an agency to the tycono.ai marketplace
Show wave analysis — role-by-role status, tokens, cost breakdown + orphan session detection. WHEN: user asks about cost, token usage, who is working, session analysis, or wants to check for zombie processes.
Executes bash commands
Hook triggers when Bash tool is used
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
Your company, in code.
Define your org. Give one order. Your AI team plans, builds, and learns — and they remember everything next time.
Terminal-native. Local-first. Open source.
claude plugin install tycono
You: /tycono "Build a landing page for my SaaS"
CEO → Breaks it down, dispatches CTO
CTO → Plans architecture, dispatches Engineer
Engineer → Writes code
QA → Reviews and tests
CEO → Approves final result
npx tycono
A setup wizard walks you through naming your company, picking a team, and starting your first wave.
Terraform turns .tf files into running infrastructure.
Tycono turns YAML and Markdown into a running company.
# roles/engineer/role.yaml
id: engineer
name: "Alex"
level: member
reports_to: cto
model: claude-sonnet-4
authority:
autonomous: ["Code implementation", "Bug fixes"]
requires_approval: ["Architecture changes"]
knowledge_scope:
readable: ["projects/", "architecture/"]
writable: ["projects/*/technical/"]
One file defines who the agent is, what they can do, what they know, and who they report to.
1. Install claude plugin install tycono
2. Build team Guided setup wizard — or pick an agency preset
3. Give order "Build a tower defense game"
4. Watch CEO dispatches CTO → Engineer → QA, in real time
5. Learn Knowledge persists. Next wave starts smarter.
A crypto trading research team ran 13.5 hours of autonomous AI work — zero human intervention:
| Metric | Value |
|---|---|
| Duration | 13.5 hours |
| Agent sessions | 174 |
| Hypotheses tested | 29 |
| Cost | $54 total ($2.5/hypothesis) |
| Critic role | Caught a false-positive before production |
Cost trajectory: $1,342/wave (early beta) → $4/wave (current).
| Feature | What it does |
|---|---|
| Wave Dispatch | One order cascades through the entire org hierarchy |
| Auto-Amend | Follow-up work continues in the same session — 70% cost reduction |
| Handoff Summary | Re-dispatched roles get previous session context automatically |
| Prompt Caching | Static system prompt sections cached — up to 90% input cost reduction |
| Critic Role | Built-in Devil's Advocate that challenges team conclusions |
| AKB | File-based knowledge system. Pre-K reads, Post-K writes. Compounds over time. |
| Heartbeat Watch | Real-time supervision. Redirect mid-execution if direction drifts. |
| Surface | Status | What |
|---|---|---|
| Plugin | ✅ Active | Claude Code plugin. Primary entry point. |
| TUI | ✅ Active | Terminal dashboard — org tree, agent streams, wave management. |
| Server | ✅ Active | Headless API backend. All surfaces consume this. |
| Pixel | ❄️ Frozen | Isometric pixel office UI. Kept for demos. |
claude plugin install tycono
/tycono "Build a landing page"
/tycono --agency gamedev "Build a tower defense"
npx tycono
┌── W1 Build the API ──────┬── Stream ──────────────────────────┐
│ 3 sessions │ cto → dispatch engineer │
│ ── Org Tree ── │ engineer → Write src/api/routes.ts│
│ ● CEO │ qa ▶ Running test suite... │
│ ├─ ● CTO │ │
│ │ ├─ ● engineer │ │
│ │ └─ ● qa │ │
│ └─ ○ CBO │ │
└───────────────────────────┴─────────────────────────────────────┘
Isometric pixel office — watch your AI team walk around and collaborate visually. Not actively developed. See tycono.ai for a preview.
Pre-built team configurations for common workflows.
npx claudepluginhub seongsu-kang/tycono --plugin tyconoMulti-agent team orchestration for Claude Code. Set up parallel AI agent teams with file-based planning, progress tracking, and role-based collaboration.
AI team role and worker manager for multi-agent development workflows.
Dynamically assemble expert agent teams for complex tasks using Claude Code's agent teams feature
AI Agent Team Operating System for Claude Code — persistent team management, meetings, task wall, company loop engine, and real-time dashboard
General-purpose agent superteam for Claude code - contract-gated verification loops with 7 outer-loop roles, 5-phase orchestration, task form-driven inner-loop cooperation, and structural prevention of premature exits.
Structured project planning and execution through brainstorm, spec, and build phases across three execution tiers: sequential, delegated sub-agents, and full agent teams