AI Security
How do you dockerize an MCP server written in JavaScript?
Package a Node MCP server in a container with a locked install, a non-root user and the right transport. Stdio needs an interactive run, HTTP needs auth.
Oct 18, 20264 min read
Resources
Supply Chain Security, in plain English.
Deep dives, practical guides, and incident analyses from engineers who build Safeguard. No fluff, no vendor FUD — just what you need to ship secure software.
Filter
Articles
RSS feedAI Security
How do you use AI for source code analysis safely?
AI can speed up code review and triage, but it misses and invents findings. Pair it with static scanners, test its output and limit what code leaves.
Oct 18, 20264 min read
AI Security
How do you run LLMs securely at scale? An operations framework
Secure LLM operations need an inventory, input and output controls, scoped tool access, logging and a review loop. Use NIST AI RMF and OWASP as the base.
Oct 18, 20264 min read
AI Security
How do you get visibility into AI-native systems?
Visibility into AI systems means seeing models, prompts, data sources, tools and agent actions. Build an inventory, log tool calls and watch for drift.
Oct 18, 20264 min read
AI Security
How do you build a secure MCP server?
A secure MCP server accepts only tokens issued for it, never forwards them, keeps scopes minimal and sandboxes local code. Here is the checklist.
Oct 18, 20265 min read
AI Security
How do AI models plan and run cyberattacks? A 2025 case
A published 2025 case shows AI agents doing most of an intrusion campaign. See how the attack was structured and which controls limit an agent's reach.
Oct 18, 20264 min read
AI Security
Hosted or self-hosted model for security work: how to choose
Choose by data sensitivity, control, operating capacity and evidence needs. A checklist for security teams deciding between a hosted API, a private deployment and a local model.
Oct 18, 20264 min read
AI Security
Generative AI vs predictive AI: how do they differ in risk?
Predictive AI scores or classifies from data. Generative AI creates new content. See how the two differ in use, failure modes and the controls each needs.
Oct 18, 20264 min read
AI Security
Generative AI vs predictive AI: how do the security risks differ?
Predictive AI classifies or forecasts. Generative AI produces new content. NIST treats both as attackable, with different failure modes. What to defend.
Oct 18, 20264 min read
Stay informed
Weekly insights on software supply chain security, delivered to your inbox.
Self-healing security runs on Safeguard.
Your first fix PR is minutes away.
No sales call required, even your agent can complete the purchase over MCP.