> For clean Markdown of any page, append .md to the page URL.
> For a complete documentation index, see https://developers.alephant.io/llms.txt.
> For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://developers.alephant.io/_mcp/server.

# Code Interpreter

> Enable controlled code execution while managing files, runtime environment, and security boundaries.

Code interpreter helps users and agents analyze files, run scripts, or process data. It is a high-impact capability and should be enabled only when the deployment has isolation and audit controls.

## Capability Boundary

| Area           | Guidance                                                                          |
| -------------- | --------------------------------------------------------------------------------- |
| Runtime        | Use an isolated execution environment without sensitive core-service permissions. |
| File Access    | Allow access only to uploaded files or workspace artifacts required for the task. |
| Network Access | Restrict internet, internal network, and credential access according to policy.   |
| Audit          | Record task, input, output, errors, and resource usage.                           |

## Before Configuration

* Confirm the execution environment, resource limits, and cleanup policy are available.
* Define which users, groups, or agents can use code interpreter.
* Do not expose database passwords, API keys, or production credentials to the runtime.

## Verification

* Run a low-risk script to validate execution, output, and file handling.
* Test that unauthorized file, network, or credential access is denied.
* Confirm errors do not leak internal paths, secrets, or stack details.