# Command injection via subprocess shell=True in Python — Agent PR X-Ray

> Canonical page: https://codevetter.com/xray/py-command-injection

This is an adjudicated synthetic benchmark example, not a claim about an uploaded repository.

## Source

- Corpus: codevetter-public-benchmark/py-command-injection
- Language: python
- Category: security
- Evidence state: benchmark_ground_truth

## Finding

- Severity: high
- Title: command injection
- Summary: subprocess.run is called with shell=True and a string command containing user-controlled host, allowing shell metacharacters to inject arbitrary commands.
- Source anchor: `source.py:9`

## Verification boundary

The static finding is qualified against benchmark ground truth. No exact-current executable test or audience validation is included in this example, so those stages remain unverified.

## Public product links

- [CodeVetter](https://codevetter.com/)
- [Download](https://codevetter.com/download)
- [Documentation](https://codevetter.com/docs/)
- [Source](https://github.com/Codevetter/codevetter)