AI Security
Why should AI agents use short-lived credentials, and how?
Long-lived keys outlast the task and survive any leak. Give each AI agent its own identity and tokens that expire in minutes, scoped to a single job.
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 should you handle authorization for AI agents?
Give each AI agent its own identity, short lived scoped tokens and a record of whose behalf it acts on. See OAuth patterns, delegation and approval gates.
Oct 18, 20265 min read
AI Security
What are adversarial attacks on AI and machine learning systems?
Adversarial attacks manipulate ML systems through evasion, poisoning and privacy attacks. Learn NIST's taxonomy and defensive steps for your own models.
Oct 18, 20264 min read
AI Security
What are adversarial AI attacks on machine learning models?
Adversarial attacks trick, poison or extract from machine learning models. NIST groups them by stage and goal. Here is how to read the taxonomy and defend.
Oct 18, 20264 min read
AI Security
What belongs in a policy for AI coding assistants at work?
A usable AI coding policy names approved tools, data rules, review duties and what agents may run. Keep it to one page and enforce it in tooling.
Oct 18, 20264 min read
AI Security
What is a data poisoning attack and how do you defend against it?
In a data poisoning attack an adversary inserts or alters training data to change a model's behavior. Provenance, validation and testing are the defenses.
Oct 18, 20264 min read
AI Security
How to test for LLM supply chain risk, OWASP LLM03 practical tests
Test LLM supply chain risk by checking provenance, hashes, packages, licenses and behavior. A defensive test plan on OWASP LLM03, with no attack code.
Oct 18, 20264 min read
AI Security
Can small models work as security classifiers?
Small, tuned models can screen prompts and responses cheaply, but they are a filter, not a boundary. Where they fit, how to test them and where they fail.
Oct 18, 20264 min read
AI Security
RAG security: how do poisoned documents and weak access control hurt you?
A RAG app is only as safe as the documents it retrieves and the permissions on them. Common mistakes and the controls OWASP lists to fix them.
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.