AI Security
Securing MCP Servers for AI Agents
Five CVEs in 2025 alone trace MCP tool compromise back to one root cause: unsanitized strings piped into exec(). Here's how to expose and consume MCP safely.
Jul 9, 20267 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 feedVulnerability Management
A prioritization framework for triaging security alerts at scale
Only 2.6% of CVEs tracked in 2019 saw real-world exploitation, per Kenna Security/Cyentia — yet most teams still triage by CVSS alone. Here's a better framework.
Jul 9, 20267 min read
Application Security
A methodology for testing SPAs for client-side vulnerabilities
DOM XSS, token storage, and API exposure don't show up in a server-side scan — here's a repeatable methodology for testing React, Vue, and Angular apps.
Jul 9, 20266 min read
Best Practices
The real ROI of shifting left: what early flaw detection actually saves
A 2025 data breach averages $4.44M globally and $10.22M in the US. Here's a defensible cost model for catching flaws before they ship, not after.
Jul 9, 20266 min read
AI Security
Signing and provenance standards for AI agent skill registries
Shai-Hulud infected 500+ npm packages via stolen tokens in 2025. Agent skill registries are repeating the same unsigned-artifact mistake — here's the fix.
Jul 9, 20267 min read
AI Security
Threat-modeling for AI-native applications
STRIDE has six categories from 1999. OWASP's LLM Top 10 and MITRE ATLAS's ~84 techniques show why agentic AI needs new threat-modeling columns, not a new framework.
Jul 9, 20267 min read
Supply Chain Attacks
Reconstructing the tj-actions/changed-files compromise
CVE-2025-30066 hit CISA's KEV list within 3 days: 23,000+ repos ran a poisoned GitHub Action that dumped CI secrets straight into public build logs.
Jul 9, 20266 min read
Best Practices
Verifying open source package provenance with SLSA and Sigstore
A maintainer account takeover hid a backdoor in xz-utils for years. SLSA provenance and Sigstore signing are how you catch the next one before it ships.
Jul 9, 20267 min read
AI Security
The OWASP Top 10 for LLM Applications, Explained
The OWASP LLM Top 10 is the closest thing the field has to a shared checklist for AI security. Here is what each of the ten risks actually means, in plain language, with the defenses that matter.
Jul 8, 20266 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.