AI Security
What is shadow AI and how do you find it in your company?
Shadow AI is AI use your security team has not approved or listed. Learn the data risks, how to find it without a witch hunt, and how to offer a safe path.
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
What is shadow AI and what risks does it create for a company?
Shadow AI is staff using AI tools nobody approved. It leaks data, adds untracked agents and creates gaps. Find it, set limits and offer a safe option.
Oct 18, 20264 min read
AI Security
What are five security practices for AI code assistants like Copilot?
Treat AI code as untrusted input: review it, scan it, check packages it names, limit data it sees and restrict what agent tools can do. A practical list.
Oct 18, 20264 min read
AI Security
How Do You Secure Generative AI When Its Outputs Vary?
Generative AI answers the same prompt differently, so one passing test proves little. Use NIST and OWASP risk lists and enforce limits outside the model.
Oct 18, 20265 min read
AI Security
How do secrets end up in prompts and context windows?
Secrets reach prompts through pasted logs, env files, tool output and retrieved documents, then persist in transcripts. Reduce exposure, scan inputs and rotate quickly.
Oct 18, 20264 min read
AI Security
How do you sandbox an AI agent that runs shell commands?
Run agent commands in a disposable container or microVM with no secrets, a read-only base, limited network and resource caps, and a human gate on anything that leaves it.
Oct 18, 20264 min read
AI Security
How do you run an MCP server in Docker safely?
Docker can sandbox a local MCP server that would otherwise run with your privileges. Use stdin, a read-only root, no capabilities and limited network.
Oct 18, 20264 min read
AI Security
How do you run an LLM locally with Docker Model Runner?
Docker Model Runner pulls open models and serves them on an OpenAI-compatible API, keeping prompts on your machine. Setup basics and the gaps to close.
Oct 18, 20264 min read
AI Security
How do you run a private mirror for Hugging Face models?
Stage vetted models internally, control the cache location, and go offline in production. Here is the setup and what a mirror does not verify.
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.