# Drop

> Updated 2026-09-22 · type: tool · category: ai-infrastructure · status: active · rev 1

Drop sandboxes untrusted programs and agent-generated code in rootless, disposable Linux home dirs via user namespaces and optional gVisor, no root needed.

- Open source: yes (Apache-2.0)
- Self-hostable: yes
- Pricing model: free
- Best for: A developer or agent-builder who runs coding-agent output or untrusted third-party programs on their own Linux machine and wants isolation without standing up containers or handing anything root.
- Last verified: 2026-09-22

- **Canonical:** https://gtmstacker.com/registry/tool/drop/
- **Source:** [github · wrr/drop](https://github.com/wrr/drop)
- **Tags:** ai-infrastructure, mcp-agents, self-hostable, sandbox, agent-security
- **Repository:** https://github.com/wrr/drop

## Is Drop open source?

Yes, Drop is open source under the Apache-2.0 license.

## How much does Drop cost?

Drop is free to use.

## Can I self-host Drop?

Yes, Drop can be self-hosted (the source is available under the Apache-2.0 license).

## Alternatives & related

- [Bastionskill](https://gtmstacker.com/registry/tool/bastionskill/)
- [Talos](https://gtmstacker.com/registry/tool/talos/)


---

Apache-2.0 rootless Linux sandbox that runs untrusted programs and agent-generated code in disposable, isolated home directories using user namespaces, with optional gVisor for a stronger boundary. Access is governed by a TOML file and nothing needs root or a container runtime, so you can execute a coding agent's output or a third-party dependency without giving it your real user account.

## Provenance

- Apache-2.0 (LICENSE independently verified), ~146 stars, ~367 commits on main (active maintenance); rootless via Linux user namespaces, optional gVisor, disposable per-run home dirs, TOML access control (WebFetch 2026-09-22, github.com/wrr/drop).
- Surfaced via the 2026-09-22 studio pull ("Drop — a rootless Linux sandbox with gVisor support," Show HN).
- Curated from the GTM Stacker signal registry (2026-09-22 pass); license independently verified 2026-09-22.

## Why it matters for a GTM stack

Once agents start writing and running code, "just run it" becomes a security decision. The usual isolation answers — Docker, microVMs — are heavier than a lot of teams want for the simple job of running one agent's output safely. Drop is the lightweight middle: rootless user-namespace isolation, a throwaway home directory per run, and a TOML file that says what the sandboxed process may touch, with gVisor available when you want a firmer wall. For anyone piping coding-agent output or untrusted dependencies through their own machine, it is a low-ceremony way to contain the blast radius, and it is Apache-2.0. The honest read: it is Linux-only, local, and modestly sized — treat it as a sane isolation habit for dev workflows, not a hardened multi-tenant runtime, and turn on gVisor if the boundary needs to hold.