k9 Security
  • Use Cases
    • DevOps
    • Enterprise Security
    • MDR & SOC
    • OEM Integrations
  • Demo
  • Pricing
  • Docs
  • Contact
  • About
  • Blog
  • Sign up
Select Page

Why coding agents need help triaging dependency alerts

by Stephen Kuenzli | Sep 1, 2026 | Development, Security

“Why not just prompt Claude Code or Copilot to triage dependency alerts per our security policy?” is the right question. A coding agent can read your code, trace a call path, and reason about whether a vulnerable function is reachable. That is the hardest...

How to Prioritize CVEs by Risk, Not Severity

by Stephen Kuenzli | Jul 20, 2026 | AI, Development, Security

In 24 hours we scored the same CVE three times against three different projects, and it came back DEFER on one and SCHEDULE on the other two. Every call was right. The scanner’s severity label read “high” all three times. What changed was the asset...

Your MCP server on AgentCore Runtime fails its health check at `/mcp/`. Here’s how to fix it.

by Stephen Kuenzli | Jun 2, 2026 | Development

We deployed the k9 MCP server to AWS Bedrock AgentCore Runtime. Tool calls worked for the first few minutes of every session, then started failing with MCP error -32010: Runtime health check failed or timed out. Local tests stayed green throughout. The fix is one...

Recent Posts

  • EPSS vs KEV: the 2% score on a ransomware CVE
  • Why coding agents need help triaging dependency alerts
  • How to Prioritize CVEs by Risk, Not Severity
  • Your MCP server on AgentCore Runtime fails its health check at `/mcp/`. Here’s how to fix it.
  • Severity is no longer a triage input. Risk scoring you own is.

Recent Comments

    Copyright 2026 K9 Security Inc.