AI Security
What is content exclusion for AI coding assistants, and its limits?
Content exclusion tells an AI coding assistant which files it must not read. Learn what it covers, who configures it, and the gaps to plan around.
Oct 18, 20263 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 an agentic coding tool, and how is it different from autocomplete?
Autocomplete suggests code you accept. An agentic coding tool plans, edits files, runs commands and can open pull requests. The security risks differ.
Oct 18, 20263 min read
AI Security
What is AI agent identity and why does it matter?
AI agent identity means each agent is a named, authenticated entity with its own limited access, so its actions can be authorized, audited and revoked.
Oct 18, 20264 min read
AI Security
What is AI agent governance? A plain definition and first steps
AI agent governance is the set of rules, owners and records that decide which agents may act, with what access, and who can prove it. Start with five.
Oct 18, 20264 min read
AI Security
What is a dataset card and which fields matter for supply chain review?
A dataset card is the README that documents a dataset. Learn which fields, license, source and bias notes, to read and record before training on it.
Oct 18, 20264 min read
AI Security
What does trust_remote_code=True do when loading a model?
trust_remote_code=True lets a Transformers model run Python code from the model repository. Learn the risk, how to pin a revision, and when to say no.
Oct 18, 20263 min read
AI Security
What does MCP say about audit logging, and what is left to you?
The MCP spec says clients should log tool usage and servers should log scope elevations. The rest of audit logging is your job. Here is the gap and a fix.
Oct 18, 20264 min read
AI Security
What can analyst reports miss about AI agents and MCP security?
Fixed criteria can lag fast-moving areas like AI agents and MCP. Learn how to check whether a report covers them and how to test the gaps yourself.
Oct 18, 20263 min read
AI Security
What are the main security risks of an MCP server?
The main MCP server risks are confused deputy, token passthrough, SSRF, session hijacking, local server compromise and broad scopes. Each has a known fix.
Oct 18, 20265 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.