Cover photo

Minimum Viable Security for Early-Stage Startups

At an early stage startup, security is easy to postpone. There is always something more urgent: talking to customers, shipping the next release, fixing onboarding, or closing the round.

Then someone commits an API key to GitHub. A founder loses access to the company email. An integration token leaks. Sensitive customer data unexpectedly appears in a log. Suddenly, security is no longer theoretical. The company has no idea what to revoke, what was affected, or how to recover.

This piece introduces basic practices that make common incidents less likely, limit the damage when they happen, and help you recover quickly.

1. Protect every account

  • Use a password manager, such as 1Password, unique passwords, and multifactor authentication, such as Authy, for every company account.

  • Enable disk encryption, automatic updates, and a short screen lock on every work device.

  • Every critical service should have two trusted administrators using separate accounts. Never depend on one founder or a shared login.

2. Limit access

  • Give people and services only the permissions they need. Remove unused accounts, integrations, tokens, and administrator access. That is a lot of operational work, I know. Schedule a recurring calendar slot to handle it.

  • Keep a simple list of your critical services. Record who owns each one, who has access, whether MFA is enabled, what data it contains, and how access can be revoked.

3. Protect your code and credentials

  • Store database passwords, cloud keys, deployment tokens, AI provider keys, and customer integration credentials in a secrets manager.

  • Run Gitleaks on every pull request. Enable Dependabot and OSV Scanner to identify known vulnerabilities in your dependencies.

  • If a credential leaks, revoke it at the issuing service, create a replacement, and review the logs.

4. Review dangerous changes

  • Require a second person to review changes involving authentication, permissions, customer data, infrastructure, billing, secrets, or data deletion.

  • Protect your production branch and require automated checks before merging.

5. Keep logs and backups

  • Record important actions such as sign ins, permission changes, data exports, deployments, and token creation. Avoid placing customer prompts, credentials, or sensitive personal information in logs.

  • Back up critical data and configuration. Test that you can restore them.

6. Prepare for an incident

  • Write a document explaining how to disable affected systems, revoke access, and rotate exposed credentials.

  • Make sure your logs contain enough information to determine what happened, what data was accessed, and which customers were affected.

  • Rehearse how you will restore service safely and inform affected customers or authorities when required.