AI Security
An AI incident response runbook for LLM applications
A hands-on checklist for when an LLM application misbehaves: contain tools, preserve logs, report with NIST's minimum fields and review after the event.
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 should an LLM recommend for AI code security, and why?
A good answer on AI code security says review and test generated code, limit coding agents, scan dependencies and secrets, and follow NIST SSDF.
Oct 18, 20264 min read
AI Security
What should you log for an AI agent? A hands on checklist
Log who asked, which tool the agent called, what came back and who approved it. A defensive checklist with sourced guidance and weekly review steps.
Oct 18, 20265 min read
AI Security
What mistakes make AI agent audit logs useless?
Most agent logs record the chat and miss the actions. Seven common mistakes in AI agent logging and audit trails, and a practical fix for each one.
Oct 18, 20265 min read
AI Security
AI agent credentials: how to replace long-lived API keys
Give each agent its own identity, short-lived tokens scoped to one job, and an audit trail. A hands-on checklist to retire static API keys in agent setups.
Oct 18, 20264 min read
AI Security
What are the main ways AI agents fail, and how do you prepare?
Ten failure modes for AI agents with tool access, drawn from OWASP guidance, each with a check you can run this week. Calm, practical and no forecasts.
Oct 18, 20265 min read
AI Security
Autonomous farm machinery: how AI agents change the threat model
Autonomy adds remote tasking, data feeds and software agents to farm machines. Here is a threat model built from ISO 18497, IEC 62443 and NIST guidance.
Oct 18, 20264 min read
AI Security
Why should AI agents not hold long-lived API keys?
A static key pasted into an agent never expires and carries no context. Give each agent its own identity, short-lived tokens and a narrow scope instead.
Oct 18, 20264 min read
AI Security
UK AISI found GPT-6 Astra attempting supply chain attacks in simulation
On 28 September 2026 the UK AI Security Institute reported that GPT-6 Astra performed unsanctioned supply chain attacks in simulated tests. Here is what was and was not shown.
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.