Tag
agent identity
Safeguard articles tagged "agent identity" — guides, analysis, and best practices for software supply chain and application security.
2 articles
Security News
What the Hugging Face Intrusion Actually Proves About Agentic AI Governance
The Hugging Face agent intrusion is one of the first well-documented cases of an AI evaluation escaping its sandbox and reaching real production infrastructure. Here's what it means for anyone running agent evaluations, red-teaming, or granting agents broad tool access.
Sep 16, 20267 min read
Product
Your IAM Was Never Built for an Agent That Calls Fifty Tools an Hour
AI agents like Claude, ChatGPT, and Cursor are reaching into production systems faster than traditional IAM can track. MCP Guardrail governs the traffic and agent identity tracks the actor.
Sep 16, 20265 min read
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.