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.
Pods allow all traffic by default. How to isolate agent sandbox pods with a default deny egress policy, a DNS exception, and a stronger runtime class.
AI assistants that create documents, spreadsheets and slides add risks around content, storage and sharing. Vendor-neutral questions to ask before buying.
A five step method to list every scheduled task, routine and automation in your AI assistants, with where each product keeps them and what to record.
Excessive agency means a coding agent has more tools, permissions, or autonomy than its task needs. Here is how to spot it and trim it in your team.
Five common myths about AI coding agent security, from approval prompts to private repos, with the facts that should shape how you set your controls.
Start with no outbound network, then add only the registries and hosts a build needs. A step by step approach to a default deny egress policy for agents.
Admin and environment switches that exist today for routines, Cowork, session cron and event triggered tasks, and what each one does not cover.
Not without checking. Treat a chatbot's KEV, EPSS or OSV number as a lead, then verify it against the source file, its date and your own count.
Agent session transcripts can hold code, command output, and credentials. How sharing settings, visibility, and retention affect what others can see.
Weekly insights on software supply chain security, delivered to your inbox.
Your first fix PR is minutes away.
No sales call required, even your agent can complete the purchase over MCP.
