/output-workflow-run

name: output-workflow-run
description: Execute an Output SDK workflow synchronously and wait for the result. Use when running a workflow and needing immediate results, testing workflow execution, or getting the output directly in the terminal.
allowed-tools: [Bash, Read, Write]

Run Workflow Synchronously

Overview

This skill executes a workflow synchronously, meaning the command waits for the workflow to complete and returns the result directly. This is ideal for testing, quick executions, and when you need immediate feedback.

When to Use This Skill

• Testing a workflow during development
• Running a workflow and needing the result immediately
• Quick one-off workflow executions
• Debugging by re-running a workflow with different inputs
• When you don't need to monitor the workflow separately

When to Use Async Instead

Consider using `npx output workflow start` (async) when:

• The workflow takes a long time (minutes to hours)
• You need to run multiple workflows in parallel
• You want to disconnect and check results later
• You need to monitor progress separately

Instructions

Basic Syntax

npx output workflow run <workflowName> --input '<json-input>'
npx output workflow run <workflowName> --input <path-to-json-file>

The `--input` flag is required when the workflow expects input data.

Input Methods

1. Inline JSON

Pass JSON directly on the command line:

npx output workflow run example --input '{"question": "who really is ada lovelace?"}'

2. File Path (Recommended)

Reference a JSON file containing the input:

npx output workflow run simple --input src/simple/scenarios/question_ada_lovelace.json

This is the recommended approach because:

• Input is version controlled and reproducible
• Complex inputs are easier to read and edit
• Scenarios can be shared and reused

Scenario Folder Pattern (Best Practice)

Workflows typically have a `scenarios/` folder containing test inputs:

src/
  my_workflow/
    workflow.ts
    steps.ts
    scenarios/
      basic_test.json
      edge_case_empty.json
      large_payload.json

**Best practice workflow:**

1. Create a scenario file with your input:

   # Create scenarios folder if it doesn't exist
   mkdir -p src/my_workflow/scenarios

2. Write your input to a scenario file:

   // src/my_workflow/scenarios/test_user.json
   {
     "userId": "123",
     "options": {
       "verbose": true
     }
   }

3. Run the workflow referencing the scenario:

   npx output workflow run my_workflow --input src/my_workflow/scenarios/test_user.json

Input Examples

# Inline JSON - simple object
npx output workflow run my-workflow --input '{"userId": "123"}'

# Inline JSON - complex nested input
npx output workflow run data-pipeline --input '{"source": "api", "options": {"limit": 100}}'

# File path - reference a scenario file
npx output workflow run simple --input src/simple/scenarios/basic.json

# File path - relative to current directory
npx output workflow run batch-processor --input ./test_inputs/batch1.json

# No input (only if workflow doesn't require it)
npx output workflow run health-check

Understanding the Output

The command returns the workflow result directly to stdout.

Success Output

The workflow's return value is displayed, typically as JSON.

Error Output

If the workflow fails, you'll see:

• Error message
• The workflow ID (for further debugging)
• Suggestion to use `npx output workflow debug` for details

Examples

**Scenario**: Test a workflow with a scenario file

# First, look for existing scenarios
ls src/simple/scenarios/

# Run using a scenario file
npx output workflow run simple --input src/simple/scenarios/basic_sum.json

# Output:
# { "sum": 6, "count": 3 }

**Scenario**: Create and run a new test scenario

# Create a scenario file
cat > src/my_workflow/scenarios/test_case_1.json << 'EOF'
{
  "question": "What is the capital of France?",
  "context": "geography"
}
EOF

# Run the workflow
npx output workflow run my_workflow --input src/my_workflow/scenarios/test_case_1.json

**Scenario**: Quick inline test during development

npx output workflow run example --input '{"question": "explain quantum computing"}'

**Scenario**: Re-run a workflow with different input for debugging

# First attempt with scenario file
npx output workflow run process-data --input src/process_data/scenarios/user_abc.json
# Error occurs

# Create a new scenario to isolate the issue
cat > src/process_data/scenarios/debug_minimal.json << 'EOF'
{"id": "test", "debug": true}
EOF

npx output workflow run process-data --input src/process_data/scenarios/debug_minimal.json

**Scenario**: Capture output for further processing

# Save result to a file
npx output workflow run generate-report --input src/generate_report/scenarios/jan_2024.json > report.json

# Pipe to jq for processing
npx output workflow run get-users --input src/get_users/scenarios/active.json | jq '.users[].name'

Error Handling

Common Errors

| Error | Cause | Solution | |-------|-------|----------| | "Workflow not found" | Workflow name is incorrect | Check with `npx output workflow list` | | "Invalid input" | JSON doesn't match schema | Verify input matches workflow's inputSchema | | "Parse error" | Malformed JSON or file not found | Check JSON syntax or file path | | "Timeout" | Workflow took too long | Use async execution for long workflows |

Getting More Details on Failures

When a workflow fails, the output includes the workflow ID. Use it to get the full trace:

npx output workflow run my-workflow --input src/my_workflow/scenarios/test.json
# Output: Workflow failed. ID: abc123xyz

npx output workflow debug abc123xyz --json

Input Schema Tips

1. **Check the schema first**: Look at the workflow's `inputSchema` in the code 2. **Use scenario fi