Skip to main content

Getting Started with Agents

Set up your first AI agent to interact with DaoFlow.

Step 1: Create an Agent Principal​

In the DaoFlow dashboard, go to Settings → Users → Add Principal and select Agent.

Step 2: Generate a Token​

Create an API token for the agent with appropriate scopes:

# Read-only token (safest starting point)
Scopes: server:read, deploy:read, service:read, logs:read, events:read

# Deploy-capable token
Scopes: + deploy:start, deploy:rollback, env:read

Step 3: Configure Your AI Tool​

Cursor / Windsurf / Cline​

Add to your agent instructions:

You have access to DaoFlow for deployment management.
Use the following CLI commands:
- daoflow status --json          # Check infrastructure health
- daoflow plan --service svc_my_app --json            # Preview a service deployment
- daoflow deploy --service svc_my_app --dry-run --json # Preview a deployment
- daoflow deploy --service svc_my_app --yes --json   # Execute a deployment
- daoflow logs --deployment dep_abc123 --json        # View deployment logs
- daoflow doctor --json         # Diagnose issues

Always use --json for structured output.
Always use --dry-run before --yes.

GitHub Actions / CI​

- name: Deploy via DaoFlow
  env:
    DAOFLOW_TOKEN: ${{ secrets.DAOFLOW_TOKEN }}
  run: |
    daoflow login --url ${{ vars.DAOFLOW_URL }} --token $DAOFLOW_TOKEN
    daoflow deploy --service svc_my_app --yes --json

Step 4: Verify​

daoflow whoami --json
daoflow capabilities --json
daoflow status --json