# YAML Deserialization via yaml.load on Untrusted Input

Language: Python
Severity: Critical
CWE: CWE-502

## Source
8

## Flow
8-9

## Sink
9

## Vulnerable Code
```python
import yaml
from flask import Flask, request, jsonify

app = Flask(__name__)

@app.route('/iot/device/provision', methods=['POST'])
def provision_iot_device():
    device_manifest = request.data.decode('utf-8')
    device_config = yaml.load(device_manifest, Loader=yaml.Loader)
    device_id = device_config.get('device_id', 'unknown')
    firmware_ver = device_config.get('firmware_version', '1.0.0')
    capabilities = device_config.get('capabilities', [])
    print(f"Provisioning device {device_id} with firmware {firmware_ver}")
    return jsonify({'status': 'provisioned', 'device_id': device_id, 'capabilities': capabilities})
```

## Explanation

The application uses yaml.load() with yaml.Loader on untrusted user input from request.data without validation. This allows arbitrary Python object instantiation during YAML deserialization, enabling remote code execution through specially crafted YAML payloads containing malicious Python objects.

## Remediation

The fix replaces yaml.load() with yaml.safe_load(), which only deserializes standard YAML tags (strings, numbers, lists, dicts) and prevents instantiation of arbitrary Python objects. Additionally, a type check ensures the parsed YAML is a dictionary before accessing its keys, providing defense against unexpected input formats.

## Secure Code
```python
import yaml
from flask import Flask, request, jsonify

app = Flask(__name__)

@app.route('/iot/device/provision', methods=['POST'])
def provision_iot_device():
    device_manifest = request.data.decode('utf-8')
    device_config = yaml.safe_load(device_manifest)
    if not isinstance(device_config, dict):
        return jsonify({'status': 'error', 'message': 'Invalid device manifest format'}), 400
    device_id = device_config.get('device_id', 'unknown')
    firmware_ver = device_config.get('firmware_version', '1.0.0')
    capabilities = device_config.get('capabilities', [])
    print(f"Provisioning device {device_id} with firmware {firmware_ver}")
    return jsonify({'status': 'provisioned', 'device_id': device_id, 'capabilities': capabilities})
```
