Landstrip
OS-level AI command sandboxing for OpenCode with Landstrip
87
+11 in 30 days
2,783
649 in 7 days
63.0
Multi-signal model
5 hours ago
2026-10-21
Install and configure
opencode.jsonWrites to this project's opencode.json — applies to this repository only.
opencode.json
{
"$schema": "https://opencode.ai/config.json",
"plugin": ["opencode-landstrip@0.19.20"]
}Writes to ~/.config/opencode/opencode.json — applies to every project.
~/.config/opencode/opencode.json
{
"$schema": "https://opencode.ai/config.json",
"plugin": ["opencode-landstrip@0.19.20"]
}If you want to modify the plugin locally, install it into the project and reference the local path.
shell
pnpm add -D opencode-landstripOpenCode loads npm dependencies through its embedded runtime on startup and caches them locally — no manual global install needed.
landstrip runs commands in an OS-level sandbox using Landlock on Linux,
Seatbelt on macOS, and AppContainer or restricted users on Windows.
Quick start
Install:
npm install --save-dev @landstrip/landstrip-api
On Linux/macOS, save this as policy.json. Windows also needs explicit read
grants for the executable and its dependencies.
{
"filesystem": {
"allowWrite": ["."],
"denyWrite": ["**/.env", "**/*.pem"],
"denyRead": ["~/.ssh"],
"allowRead": ["~/.ssh/config"]
},
"network": {
"allowNetwork": false,
"allowLocalBinding": false
}
}
npx landstrip run -p policy.json -- cargo test
npx landstrip check -p policy.json
npx landstrip capabilities
npx landstrip doctor
See landstrip(1) for CLI options, policy rules, and platform limits.
On Linux, commands always run in new user, PID, and mount namespaces with a
private /proc. Landlock and seccomp still enforce filesystem and network
policy. Execution fails closed when namespaces are unavailable; check,
doctor and capabilities report the kernel error.
Integrations
License
- Native sandbox: LGPL-2.1-or-later.
- Node.js API and agent extensions: Apache-2.0; see each package's
LICENSE.
Similar plugins
Seatbelt
opencode-seatbelt
macOS Seatbelt (kernel sandbox) for OpenCode. Stops coding agents from reading secrets, enforced by the OS, not by command matching.
V2 Security
opencode-v2-security
Autonomous execution-boundary security for OpenCode V2: scoped assessments, one-call authorization, bypass leases, and session read/write permission ceilings.
Goal
@prevalentware/opencode-goal-plugin
OpenCode goal plugin that adds Codex-style long-running goal mode, /goal commands, persistence, and TUI status for AI coding agents.