README.md

Aido Examples

This directory contains example sandbox configurations and tasks to help you get started with Aido.

---

Sandbox Configurations

sandbox.readonly.yaml

A minimal read-only sandbox configuration for tasks that only need to analyze or inspect data.

Use cases:

• Log analysis
• Data inspection
• Report generation
• Code review

Run:

aido sandbox up -f sandbox.readonly.yaml
aido run tasks/analyze_logs.py

---

sandbox.mutate.yaml

A sandbox configuration with write permissions for tasks that need to modify files.

⚠️ CAUTION: Requires --dangerous-allow-mutations flag

Use cases:

• File transformation
• Code generation
• Build processes
• Data migration

Run:

aido sandbox up -f sandbox.mutate.yaml
aido run tasks/transform.ts --dangerous-allow-mutations

---

sandbox.cloud.yaml

A cloud-aware sandbox configuration for tasks that interact with AWS infrastructure.

Note: Cloud permissions are a stub in v0.1 and not yet functional.

Use cases:

• Infrastructure checks
• Resource auditing
• Configuration validation
• Cost analysis

---

Example Tasks

tasks/hello.py

A simple Python "Hello World" task that demonstrates:

• Reading from inputs (if available)
• Writing to outputs
• Logging to stdout
• Basic file operations

Run:

aido sandbox up
aido run examples/tasks/hello.py
aido logs
cat .aido/sandboxes/latest/outputs/hello.txt

---

tasks/hello.ts

A Node.js/TypeScript version of the Hello World task.

Run:

aido sandbox up -f examples/sandbox.mutate.yaml
aido run examples/tasks/hello.ts

---

tasks/analyze_logs.py

A more realistic example that:

• Reads multiple log files from inputs
• Analyzes log levels and errors
• Generates JSON and Markdown reports
• Writes multiple outputs

Try it:

# Create sample log files
mkdir -p inputs
echo "[INFO] Application started" > inputs/app.log
echo "[ERROR] Database connection failed" >> inputs/app.log
echo "[WARN] Memory usage high" >> inputs/app.log

# Run analysis
aido sandbox up -f examples/sandbox.readonly.yaml
aido run examples/tasks/analyze_logs.py

# View results
cat .aido/sandboxes/latest/outputs/report.md
cat .aido/sandboxes/latest/outputs/analysis.json

---

Creating Your Own Tasks

Basic Task Template (Python)

#!/usr/bin/env python3
import sys
from pathlib import Path

def main():
    # Read inputs
    inputs_dir = Path("/inputs")
    # ... process inputs ...

    # Write outputs
    outputs_dir = Path("/outputs")
    outputs_dir.mkdir(exist_ok=True)

    output_file = outputs_dir / "result.txt"
    output_file.write_text("Task complete!")

    print("✓ Task completed successfully")
    return 0

if __name__ == "__main__":
    sys.exit(main())

Basic Task Template (TypeScript/Node)

#!/usr/bin/env node
import { writeFileSync, mkdirSync } from 'fs';
import { join } from 'path';

function main(): number {
  // Read inputs from /inputs

  // Write outputs
  const outputsDir = '/outputs';
  mkdirSync(outputsDir, { recursive: true });

  writeFileSync(
    join(outputsDir, 'result.txt'),
    'Task complete!',
    'utf-8'
  );

  console.log('✓ Task completed successfully');
  return 0;
}

process.exit(main());

---

Directory Structure in Sandboxes

When your task runs inside a sandbox, it has access to:

/inputs/          # Read-only inputs (mounted from your workspace)
/work/            # Working directory (writable, ephemeral)
/outputs/         # Output directory (writable, persisted)

Tips:

• Always write results to /outputs/
• Use /work/ for temporary files
• Never try to write to /inputs/ (read-only)

---

Next Steps

1. Try running the example tasks
2. Modify them to suit your needs
3. Create your own sandbox.yaml configurations
4. Build custom tasks for your workflows

---

Need Help?

• Read SANDBOX_SCHEMA.md for full schema documentation
• Check TECH_SPEC.md for implementation details
• Open an issue if something's unclear

---

Remember: Trust comes from constraints. Start with readonly, graduate to mutations.