VPN Works: Security Without Hiring a Security Team
Bootstrapped founders know that hiring a security specialist is out of reach. You need your infrastructure to be safe, but you also need to keep payroll lean. The tradeoff has always been uncomfortable: either you build products and hope nothing breaks, or you slow down to implement security controls that you’re not sure you even need.
Now add coding agents to the mix. An agent can cut your feature development time in half, but you have no idea what it’s connecting to. It reads code from your repo, issues from your tracker, snippets from web pages. One bad instruction planted in text and the agent could send your deploy credentials to a server you don’t control. You want to use the agent, but the security risk feels real.
VPN Works solves this without requiring security expertise or hiring staff. The tool seals an agent in a Linux network namespace whose only way out is through a policy that you drafted. A script that tries to ignore proxy settings finds no route. Every connection the agent makes is logged and checked against rules. The agent can work freely within the policy, and security is built in.
Day to day it comes down to four commands. run starts the agent with a given policy. trace logs everything the agent connects to. guard enforces a policy and blocks what’s outside it. learn drafts a policy from what the agent actually did.
The learn command is the bootstrapper’s win. You run the agent once doing normal work. You log everything it connects to. You read that log, see that the agent pinged GitHub to read issues and npm to fetch packages and DockerHub to pull images, strike anything obviously wrong, and lock in the rest. Under that policy, the agent works freely. A hostname that isn’t allowed is never even looked up by DNS, so DNS itself can’t leak data.
For bootstrapped founders running agents on their own machines, this is the security model that makes sense. The agent gets autonomy within bounds. The bounds are readable because you drafted them from watching the agent work normally. Every connection is logged, so if something does go wrong, you have the evidence. Security is real, not theater, and it didn’t require hiring anyone.
The policy engine runs on your machine, so you iterate on policy without waiting for anyone’s approval. You see the agent tried to connect somewhere, edit the policy to allow or block it, and test the change against old logs to see what would have happened. The feedback loop is minutes.
There’s a second engine. Scope brings the same idea to the company VPN. It learns who actually uses which systems inside your network and drafts least-privilege rules for the gateway, so a stolen login doesn’t open the whole office. Scope is for when you grow large enough to have a gateway; VPN Works is for now.
Both tools are Alphas, tested on Linux. The live demo replays real runs of an agent trying to leak credentials, and you can edit the policy against the real engine, compiled to WebAssembly. How It Works covers the sandbox in detail.
For bootstrapped founders, the appeal is self-reliance. You don’t hire a security specialist. You don’t pay for a managed security service. You draft a policy from watching your agent work, enforce it on your machine, and you have proof of what’s happening. Security is under your control, documented in a format you understand, and costs nothing.
That’s how bootstrapped companies stay safe: they use tools that multiply their ability to manage security instead of hiring specialists. VPN Works lets you run agents safely without becoming a security expert.