Skip to content

First Steps

Run your first AI workflow and see results in under 5 minutes.

Your First Workflow

Let's run a security audit on some code. This workflow scans for vulnerabilities and provides actionable recommendations.

Option 1: CLI (Fastest)

# Scan your source directory
attune workflow run security-audit --path ./src

# Or scan current directory
attune workflow run security-audit --path .

Example output: 

Security Audit Results
======================
Status: completed
Findings: 3

[HIGH] SQL query vulnerable to injection
  File: src/database.py:45
  Fix: Use parameterized queries

[MEDIUM] Hardcoded API key in config
  File: src/config.py:12
  Fix: Move to environment variable

[LOW] Missing input validation
  File: src/handlers/user.py:28
  Fix: Validate user input before processing

Cost: $0.0850

Option 2: Python

from attune.workflows import SecurityAuditWorkflow
import asyncio

async def audit():
    workflow = SecurityAuditWorkflow(enable_cache=True)
    result = await workflow.execute(target_path="./src")

    print(f"Status: {result.status}")
    print(f"Found {len(result.findings)} issues:")

    for finding in result.findings:
        print(f"  [{finding.severity}] {finding.description}")

    print(f"\nCost: ${result.cost_report.total_cost:.4f}")

asyncio.run(audit())

Try More Workflows

Attune AI includes 10+ built-in workflows:

Workflow Command What It Does
Security Audit attune workflow run security-audit Find vulnerabilities
Bug Prediction attune workflow run bug-predict Predict likely bugs
Test Coverage attune workflow run test-coverage Generate missing tests
Release Prep attune workflow run release-prep Pre-release checklist
Dependency Check attune workflow run dependency-check Find outdated deps 
# List all available workflows
attune workflow list

# Get help for a specific workflow
attune workflow run security-audit --help

Understanding the Output

Every workflow returns:

Field Description
status success, partial, or failed
findings List of issues found (for analysis workflows)
cost_report API costs and cache hit rate
metadata Timing, model used, etc.

Cost Tracking

# See your usage
attune telemetry show

# See cost breakdown
attune telemetry savings --days 7

All telemetry data stays local in ~/.empathy/telemetry/.

What Just Happened?

When you ran the security audit:

  1. File scanning - The workflow read your source files
  2. Tiered analysis - Simple files used a cheap model, complex ones used a capable model
  3. Pattern matching - Known vulnerability patterns were detected
  4. Report generation - Results were formatted and returned

This is the tiered model approach - automatically routing to the right model based on task complexity.

Next Step

Now that you've run a workflow, it's time to Choose Your Path based on how you want to use the framework.

See Also

Quick Reference

# Run workflows
attune workflow run <name> --path <path>
attune workflow list

# Check status
attune telemetry show
attune telemetry savings

# Get help
attune --help
attune workflow --help