AI Agent Security

We specialize in securing AI agents that have real access to real systems. Not theoretical. We run agents in production that manage business operations, and we’ve built the tools to keep them in check.

  • Runtime egress protection (Pipelock)
  • MCP server security scanning
  • Workspace integrity monitoring
  • Prompt injection detection
  • OWASP Agentic Top 10 compliance mapping

Security Research & Consulting

We publish what we find. Recent work includes research on lateral movement in multi-agent LLM systems and analysis of credential leaks in AI agent skill marketplaces.

  • AI agent threat modeling
  • OWASP-aligned security assessments
  • Open-source security tooling
  • Skill and plugin security auditing

Infrastructure & Networking

We build and manage on-prem infrastructure for people who want control over their stack.

  • Kubernetes (K3s) cluster design and management
  • Network architecture and security hardening
  • Automated backup and disaster recovery
  • GPU compute for AI workloads

The Other Kind of Pipe Fitting

PipeLab started in the trades. Real plumbing, residential and commercial. Turns out the mindset for diagnosing a slab leak and debugging a network issue isn’t that different. Both kinds of infrastructure break at 2 AM.

  • Residential and commercial plumbing
  • New construction and service/repair