GitHub Repository

Local control plane for running AI agents with sandboxes, approvals, guardrails, credentials, and runtime health.

56 starsTypeScript

Armorer – A secure local control plane for AI agents

Name: Armorer – A secure local control plane for AI agents
Availability: InStock
Author: cristianleo

by cristianleo·May 8, 2026·4 points·1 comment

Visit Project View on HN

AI Analysis

●●SolidShip ItSolve My Problem

Agents install their own sandbox via Docker, solving the dependency hell problem.

Strengths

•True process isolation via Docker containers prevents host machine compromise.
•Unified UI and CLI for monitoring job status across different agent types.
•Self-installation workflow where agents configure their own environment securely.

Weaknesses

•Heavy dependency on Docker Desktop running locally may limit adoption on some setups.
•Competes with emerging native sandboxing features in tools like Cursor and Claude Code.

Post Description

Hey HN,

I built Armorer because I was tired of two things: 1. The absolute "dependency hell" of setting up new AI agents (Codex, OpenClaw, etc.). 2. The security risk of giving powerful local agents broad access to my host machine.

Armorer is a secure local control plane that manages the lifecycle of your agents. It uses Docker for true process isolation and provides a unified UI/CLI for monitoring and job tracking.

One feature I’m particularly excited about: you can point an existing coding agent at the Armorer repo, and it will autonomously install and configure the entire stack for you securely.

Source: https://github.com/ArmorerLabs/Armorer Website: https://armorerlabs.com

I’d love to hear your thoughts on the architecture and how you're handling agent security locally.