Multi-agent fleet management for OpenClaw
One CLI to monitor, manage, and orchestrate your entire AI agent fleet.
📖 Read the story behind this →
Quick Start • Commands • Patterns • Configuration • Docs
You're running multiple OpenClaw gateways: a coordinator that thinks, employees that code, review, deploy, and research. Fleet is the operational layer that was missing. One CLI that gives your coordinator full visibility, control, and judgment over the entire fleet.
See what changed. Dispatch tasks. Learn which agents actually deliver, and route work to whoever you can trust. Works with any runtime.
Built for AI agents to manage AI agents. Works on any system 🦞
🔍 Visibility · Know which agents are up, which CI is red, what changed overnight. One command, full picture.
📊 Delta tracking · SITREP remembers the last run. Only shows what changed. No noise.
🔧 Zero config · fleet init detects running gateways, discovers your workspace, links itself to PATH. One command to go from clone to operational.
🧩 Modular · Each command is a separate file. Adding a new command means dropping a .sh file in lib/commands/. No monolith, no framework.
⚡ Agent native · Designed to be used by agents, not just humans. The SKILL.md teaches any OpenClaw agent to manage a fleet autonomously, install dependencies, and adapt to any environment. If bash isn't available, your agent figures out another way.
📦 Pattern library · Solo empire, dev team, research lab. Pre built configs for common setups.
- Why Fleet?
- Quick Start
- Commands
- Patterns
- Configuration
- Environment Variables
- Architecture
- For AI Agents
- Requirements
- Roadmap
- Contributing
- Support
# Install via ClawHub
clawhub install fleet
# Or via skills.sh (works with Claude Code, Codex, Cursor, Windsurf, and more)
npx skills add oguzhnatly/fleet
# Or clone directly
git clone https://github.com/oguzhnatly/fleet.git
fleet/bin/fleet init # links PATH, detects gateways, creates config
# Check your fleet
fleet agents
fleet health
fleet sitrep| Command | Description |
|---|---|
fleet health |
Health check all gateways and endpoints |
fleet agents |
Show agent fleet with live status and latency |
fleet sitrep [hours] |
Full SITREP with delta tracking |
fleet audit |
Check for misconfigurations and risks |
| Command | Description |
|---|---|
fleet ci [filter] |
GitHub CI status across all repos |
fleet skills |
List installed ClawHub skills |
| Command | Description |
|---|---|
fleet backup |
Backup gateway configs, cron jobs, auth profiles |
fleet restore |
Restore from latest backup |
fleet init |
Interactive setup with gateway detection |
See more command output examples
Agent Fleet
───────────
⬢ coordinator coordinator claude-opus-4 :48391 online 13ms
⬢ coder implementation codex :48520 online 8ms
⬢ reviewer code-review codex :48540 online 9ms
⬡ deployer deployment codex :48560 unreachable
⬢ qa quality-assurance codex :48580 online 7ms
Fleet Audit
───────────
Configuration
✅ Config file exists at ~/.fleet/config.json
✅ Config permissions: 600
✅ All agent tokens configured
✅ No placeholder tokens found
Agents
✅ All 5 agents online
✅ Main gateway healthy (:48391)
CI
✅ gh CLI available
✅ All CI green
Resources
✅ Memory usage: 43%
✅ Disk usage: 7%
Backups
✅ Last backup: 2 day(s) ago
All clear · 11 checks passed, 0 warnings
CI Status
─────────
frontend (myorg/frontend)
✅ Update homepage (main) passed 2h ago
✅ Fix footer (main) passed 4h ago
backend (myorg/backend)
❌ Add endpoint (main) failed 1h ago
✅ Fix auth (main) passed 3h ago
Fleet Health Check
──────────────────
✅ coordinator (:48391) 12ms
Endpoints
✅ website (200) 234ms
✅ api (200) 89ms
❌ docs UNREACHABLE
Services
✅ openclaw-gateway
Fleet supports any agent organization pattern. Three common ones:
One brain, many hands. The indie hacker setup.
Coordinator (Opus)
/ | \
Coder Reviewer Deployer
(Codex) (Codex) (Codex)
Team leads managing specialized developers.
Orchestrator (Opus)
/ | \
FE Lead BE Lead QA Lead
(Sonnet) (Sonnet) (Sonnet)
/ \ | |
Dev1 Dev2 Dev1 Tester
(Codex)(Codex)(Codex) (Codex)
Specialized agents for knowledge work.
Director (Opus)
/ | \ \
Scraper Analyst Writer Fact Check
(Codex) (Sonnet) (Sonnet) (Codex)
See docs/patterns.md for detailed guides and examples/ for configs.
Fleet reads ~/.fleet/config.json. Create one with fleet init or manually:
{
"workspace": "~/workspace",
"gateway": {
"port": 48391,
"name": "coordinator"
},
"agents": [
{ "name": "coder", "port": 48520, "role": "implementation", "model": "codex" },
{ "name": "reviewer", "port": 48540, "role": "code review", "model": "codex" }
],
"endpoints": [
{ "name": "website", "url": "https://myapp.com" },
{ "name": "api", "url": "https://api.myapp.com/health" }
],
"repos": [
{ "name": "frontend", "repo": "myorg/frontend" },
{ "name": "backend", "repo": "myorg/backend" }
]
}Everything is configurable. No hardcoded ports, models, or names. Your fleet, your way.
See docs/configuration.md for the full schema.
| Variable | Description | Default |
|---|---|---|
FLEET_CONFIG |
Config file path | ~/.fleet/config.json |
FLEET_WORKSPACE |
Workspace override | Config value |
FLEET_STATE_DIR |
State persistence | ~/.fleet/state |
NO_COLOR |
Disable colors | (unset) |
fleet/
├── bin/fleet # Entry point
├── lib/
│ ├── core/ # Config, output, state management
│ │ ├── config.sh # JSON config loader
│ │ ├── output.sh # Colors, formatting, HTTP helpers
│ │ └── state.sh # Delta state persistence
│ └── commands/ # One file per command
│ ├── agents.sh # Agent fleet status
│ ├── audit.sh # Misconfiguration checker
│ ├── backup.sh # Config backup/restore
│ ├── ci.sh # GitHub CI integration
│ ├── health.sh # Endpoint health checks
│ ├── init.sh # Interactive setup
│ ├── sitrep.sh # Structured status reports
│ └── skills.sh # ClawHub skill listing
├── templates/configs/ # Config templates
├── examples/ # Architecture pattern examples
│ ├── solo-empire/
│ ├── dev-team/
│ └── research-lab/
├── docs/ # Documentation
├── tests/ # Integration tests
├── SKILL.md # ClawHub agent instructions
└── .github/workflows/ # CI pipeline
Modular by design. Each command is a separate file. Add your own by dropping a .sh file in lib/commands/.
Fleet ships with a SKILL.md that any AI coding agent can read. Install it and your coordinator automatically knows how to manage the fleet:
clawhub install fleet # OpenClaw agents
npx skills add oguzhnatly/fleet # Claude Code, Codex, Cursor, Windsurf, etc.The agent reads the skill file, learns the commands, and runs health checks autonomously during heartbeat cycles.
| Dependency | Version | Notes |
|---|---|---|
| bash | 4+ | macOS ships 3.2, run brew install bash |
| python3 | 3.10+ | No pip packages needed |
| curl | any | Pre installed on most systems |
| OpenClaw | any | Gateway support required |
| gh CLI | any | Optional, for CI commands |
Fleet is being built in stages. Each version makes it more active, more intelligent, and more universal.
Visibility layer. Monitoring, delta SITREP, CI status, backup, audit. Fleet can see the entire operation.
Fleet stops being observational and becomes directive.
-
fleet task <agent> "<prompt>"— dispatch a task to any agent from the CLI -
fleet steer <agent> "<message>"— send a mid-session correction to a running agent -
fleet watch <agent>— live log tail from a specific agent session -
fleet parallel "<task>"— describe a high-level task, fleet breaks it into subtasks, assigns to the right agents, runs in parallel -
fleet log— append-only structured log of everything dispatched and received -
fleet kill <agent>— graceful session end
Fleet learns which agents actually deliver, not just which ones are alive.
-
fleet trust— trust matrix for all agents: delivery rate, rejection rate, consistency, trend -
fleet score <agent>— detailed PDR breakdown with failure patterns and trajectory - Reliability-weighted task routing for
fleet parallel(dispatch to best agent for task type, not just whoever is idle) - Trend scoring: an agent at 30% reject rate trending down ranks higher than one at 10% trending flat
PDR (Provenance-Driven Reputation): delivered / dispatched * time_modifier. Built on observable behavior from the fleet log, not self-reported health.
Fleet works with any agent on any runtime, not just OpenClaw.
- Standardized adapter interface: OpenClaw, Claude Code, Codex, generic HTTP, Docker
-
fleet adapters— list registered adapters and their status -
fleet runtime add <name> <type>— register a new runtime without editing config manually - Backward compatible: existing configs default to OpenClaw adapter
Fleet becomes an embeddable data source, not just a CLI.
-
fleet serve— start fleet as a local HTTP server - REST API:
/agents,/sitrep,/trust,/log,/task,/steer - External tools, dashboards, and CI systems can consume fleet data
- Foundation for a future cloud sync tier
Issues and PRs welcome. See CONTRIBUTING.md for guidelines.
Built by a solo operation, for solo operations.
If Fleet is useful to you, consider supporting its development:
