Security best practices

Use when designing, building, or reviewing application security.

Mental model

• Security starts with reducing attack surface, enforcing least privilege, and making unsafe states hard to reach by default.
• Treat every trust boundary explicitly: user input, APIs, files, cookies, webhooks, queues, internal services, and third-party integrations.
• Prefer simple, auditable controls over clever security mechanisms that few people can maintain.

Threat modeling

• Identify assets, attackers, entry points, trust boundaries, and likely abuse cases before choosing controls.
• Focus first on what would materially harm users or the business: account takeover, data exfiltration, privilege escalation, destructive writes, supply-chain compromise, and service abuse.
• Revisit threat models when adding new integrations, auth flows, sensitive data, or execution surfaces.

Authentication and authorization