1
0
Fork 0
forked from fun/fun
fun/web/documentation/security-and-sandboxing.md

43 lines
1.1 KiB
Markdown

---
layout: page
published: true
noToc: true
noComments: false
noDate: false
title: Fun - Security and Sandboxing
subtitle: Trust boundaries, I/O expectations, and capability restrictions.
description: Trust boundaries, I/O expectations, and capability restrictions.
permalink: /documentation/security-and-sandboxing/
lang: en
tags:
- documentation
- handbook
- installation
- usage
- introduction
- help
- guide
- howto
- docs
- specifications
- specs
- repl
---
# Security and Sandboxing
Understand the trust boundaries and how to run Fun code safely.
## Trust model
- By default, Fun code can access functionality exposed by the stdlib and any enabled extensions.
- File and network access depend on available modules and host configuration.
## Running untrusted code
- Prefer running in a container/VM with restricted filesystem and network.
- Limit available stdlib/modules by controlling `FUN_LIB_DIR` contents.
- Use OS-level sandboxing (seccomp, AppArmor, SELinux, chroot) where applicable.
## Best practices
- Avoid running as root.
- Validate and sanitize inputs at module boundaries.
- Keep your build minimal; disable unneeded extensions at compile time.