Multi-agent orchestration system for OpenCode
Senate brings multi-agent orchestration to OpenCode. Zero learning curve.
Don't learn orchestration. Just use Senate.
Senate is a multi-agent orchestration system for OpenCode. It provides:
- 19 specialized agents for different task types
- 39 workflow skills for common development tasks
- Execution modes: autopilot, ralph, ultrawork, team, ultraqa
- Planning workflows: deep-interview, ralplan, plan
- Intelligent model routing: Automatic tier selection
- State persistence: Cross-session memory and progress tracking
# Clone or download to your project
git clone https://github.com/senate/opencode-senate.git .opencode-senate
# Link to your project's .opencode directory
ln -s .opencode-senate/.opencode .opencode# Install globally
npm install -g opencode-senate
# Or use in project
npm install opencode-senate# In an OpenCode session
/ralph fix all TypeScript errors
/autopilot build a REST API for managing tasks
/ultrawork implement auth in parallel
/team 3:executor "add feature X"| Mode | Description | Use For |
|---|---|---|
| ralph | Persistent execution until verified complete | Tasks that must complete fully |
| autopilot | Full autonomous execution from idea to code | End-to-end feature development |
| ultrawork | Maximum parallel agent execution | Burst parallel fixes/refactors |
| team | N coordinated agents on shared tasks | Complex multi-step workflows |
| ultraqa | QA cycling until goal met | Testing and verification loops |
Senate includes 20 specialized agents:
- executor - Code implementation and refactoring
- architect - System design and architecture
- planner - Task sequencing and planning
- debugger - Root-cause analysis
- code-reviewer - Comprehensive code review
- security-reviewer - Security audits
- test-engineer - Test strategy and coverage
- designer - UI/UX design and interfaces
- frontend-design - Distinctive, production-grade frontend interfaces
- writer - Documentation
- analyst - Requirements analysis
- explore - Fast codebase search
- verifier - Completion verification
- tracer - Agent flow tracking
- git-master - Git operations
- critic - Plan/design challenge
- qa-tester - Interactive validation
- scientist - Research tasks
- code-simplifier - Code cleanup
- document-specialist - SDK/API docs
Activate modes with natural language:
"ralph","don't stop"→ Persistent mode"autopilot","build me"→ Full autonomous"ultrawork","ulw"→ Maximum parallelism"deep interview"→ Socratic clarification"cancel","stop"→ Cancel active modes
.opencode/
├── opencode.jsonc # Main configuration
├── AGENTS.md # Orchestration instructions
├── agent/ # 19 agent definitions
│ ├── executor.md
│ ├── architect.md
│ └── ...
├── skills/ # 39 workflow skills
│ ├── ralph/SKILL.md
│ ├── autopilot/SKILL.md
│ └── ...
├── command/ # Custom commands
├── tool/ # Custom tools
└── plugins/
└── senate-plugin.ts # Hook system plugin
Senate uses .senate/ directory for persistent state:
.senate/state/- Mode state files (JSON).senate/notepad.md- Session notes.senate/project-memory.json- Cross-session knowledge.senate/plans/- Planning documents
Add to your opencode.jsonc:
MIT
{ "plugin": ["opencode-senate"], "instructions": ["AGENTS.md"], "skills": { "paths": ["./.opencode/skills"] } }