Flow Commands

The astonish flows command manages agent flow pipelines from the command line.

Usage

astonish flows <subcommand> [flags]

Subcommands

List Flows

astonish flows list

Run a Flow

# Run a flow by name
astonish flows run <flow-name>

# Run with parameters
astonish flows run my-flow -p file=data.csv -p format=json

# Run with a specific model
astonish flows run my-flow --provider openai --model gpt-4o

# Auto-approve all tool calls
astonish flows run my-flow --auto-approve

Show Flow Structure

# Visualize a flow's node structure
astonish flows show <flow-name>

Edit a Flow

# Open flow YAML file for editing
astonish flows edit <flow-name>

Import a Flow

# Import a flow from a local YAML file
astonish flows import <path-to-flow.yaml>

Remove a Flow

astonish flows remove <flow-name>

Flow Store

Browse and install flows from configured taps:

# List available flows from all taps
astonish flows store list

# Search for flows
astonish flows store search <query>

# Install a flow from a tap
astonish flows store install <tap-name>/<flow-name>

# Uninstall a tap flow
astonish flows store uninstall <tap-name>/<flow-name>

# Update all tap manifests
astonish flows store update

Flags for run

Flag	Short	Description
--provider		AI provider to use
--model		Model name
-p		Parameter in key=value format (repeatable)
--auto-approve		Auto-approve all tool executions
--browser		Launch with embedded web browser UI
--port		Port for web server (with --browser, default: 8080)
--debug		Enable debug mode

Scheduling

Flows can be scheduled for recurring execution. Ask the agent to schedule a flow, or manage existing schedules with the scheduler.