A portable, publishable OpenClaw starter for a multi-agent operating model with a Feishu-driven project layer.
It is designed for operators who want:
- repeatable OpenClaw setup on new servers
- role-based agents instead of one overloaded assistant
- Git-managed templates and versioned operational defaults
- Feishu docs / bitable as the live project operating layer
This starter fits best if you:
- run OpenClaw on your own server or workstation
- want specialist agents such as
arch,dev,tester,devops - use Feishu for project coordination
- want a repo you can reuse across machines and publish safely
graph TD
A[GitHub Repo Layer] --> B[Runtime Layer]
B --> C[Feishu Operations Layer]
A --> A1[Templates]
A --> A2[Configs]
A --> A3[Scripts]
A --> A4[Docs]
B --> B1[openclaw.json]
B --> B2[Workspaces]
B --> B3[Secrets and Auth]
B --> B4[Systemd Service]
C --> C1[Project Homepages]
C --> C2[PM Bitable]
C --> C3[Risks and Decisions]
C --> C4[Release Records]
Rule: GitHub manages the blueprint. Feishu manages the live operating surface.
- config templates for multi-agent OpenClaw
- workspace templates for specialist agents
- project templates for repeatable project kickoff
- systemd templates for user-level deployment
- bootstrap and validation scripts for initialization and release checks
- operational docs for publishing, migration, and role boundaries
openclaw-stack/
├── README.md
├── LICENSE
├── CONTRIBUTING.md
├── SECURITY.md
├── .env.example
├── .gitignore
├── configs/
│ └── openclaw.multi-agent.json5
├── docs/
│ ├── repository-guide.md
│ ├── agent-operating-model.md
│ ├── quickstart-feishu.md
│ ├── public-release-checklist.md
│ └── limitations.md
├── systemd/
│ ├── openclaw-gateway.service
│ └── env.example
├── workspaces/
│ ├── main/
│ ├── arch/
│ ├── dev/
│ ├── solidity/
│ ├── tester/
│ ├── devops/
│ └── intel/
├── project-template/
│ ├── overview.md
│ ├── status.md
│ ├── decisions.md
│ ├── todo.md
│ ├── prd/
│ ├── architecture/
│ ├── backend/
│ ├── frontend/
│ ├── contracts/
│ ├── testing/
│ ├── deploy/
│ └── research/
└── scripts/
├── bootstrap.sh
├── init-workspaces.sh
├── install-systemd.sh
├── sync-config.sh
├── first-run-check.sh
└── validate-public-release.sh
main: intake, routing, synthesisarch: requirements / architecture reviewdev: frontend/backend implementationsolidity: contracts / on-chain safetytester: acceptance / regression / edge casesdevops: AWS delivery / monitoring / rollback / costintel: AI + Web3 research
See:
docs/agent-operating-model.mddocs/quickstart-feishu.md
git clone <your-repo-url>
cd openclaw-stackcp .env.example .env
# fill your provider keys and token placeholdersbash scripts/first-run-check.shbash scripts/bootstrap.shbash scripts/install-systemd.sh
systemctl --user restart openclaw-gatewayBefore making the repository public:
bash scripts/validate-public-release.shThen review:
- placeholder values vs real values
- channel defaults and surface exposure
- docs that may still assume your private environment
- service assumptions for your target audience
See docs/public-release-checklist.md for the final pass.
Do not commit:
.env- real API keys
- real
/etc/openclaw/env - auth/session artifacts
- provider secrets from runtime files
Use placeholders everywhere in public templates.
See SECURITY.md for details.
This repo is intended to be reusable across machines, but you should still review:
- provider availability
- channel configuration
- file paths
- systemd assumptions
- user-level vs system-level service preferences
This is a strong starting point, not a turnkey production guarantee.
See docs/limitations.md for what is intentionally left to the operator.
- add environment-variable rendering for config templates
- add a first-run healthcheck against a real OpenClaw install
- add example Feishu project homepage docs
- add CI checks for public-release validation
- add screenshots for public README polish
MIT