Inktrace: Wiz for AI Agents

🐙 Solving enterprise blind spots in AI agent ecosystems. Monitoring agent activity in real-time, detecting threats and compliance violations.

Comment

🐙 Inktrace: Agent Security Intelligence from the Deep

💡 The Problem We Solve

There is a tremendous focus in many organizations to build or deploy AI agents for various use cases. But here's the terrifying reality: there's no way to monitor what these agents are actually doing.

When regulatory bodies, auditors or potential client asks "Are your AI agents compliant with Australian safety standards?" and you can't answer, the Million dollar deal dies instantly.

🎯 What Inktrace Does

Inktrace is the attempt at creating AI agent security observatory using octopus-inspired distributed intelligence. Our specialized security tentacles monitor multi-agent ecosystems in real-time using Google's A2A protocol:

  • Real-time threat detection with instant dashboard updates
  • Australian AI Safety Guardrails compliance automation
  • Enterprise security posture with live monitoring
  • Regulatory compliance preventing deployment blockers

We don't just want to catch malicious agents—we want to catch the compliance gaps that actually shut down AI projects.

🛠 How We Built It

Built on Google's Agent2Agent (A2A) protocol with production-ready architecture:

  • 8-Tentacle System: Each specializes in security domains (Identity, Data Protection, Behavioral Intelligence, Compliance, etc.)
  • Real-time WebSocket dashboard for live threat visualization
  • Australian AI Safety Standard integration with BigQuery compliance database
  • Multi-agent ecosystem demonstrating distributed security intelligence

🏆 What We've Accomplished

  • Basic first version of agent security platform built on open A2A standard
  • Real Australian government compliance (AI Safety Standard 2024)
  • Working multi-agent demonstrations including malicious agent detection

Demo Ready: Watch malicious agents detected in real-time, compliance violations flagged instantly, and enterprise security posture updated live.

📚 What We Learned

The real enterprise killer isn't technical threats—it's regulatory compliance failure. In our regulated industries, we've seen massive deals die from missing compliance documentation. AI agents face the same risk: when enterprise buyers ask about governance standards and you can't answer, the sale is over.

Australian organizations need governance, compliance, risk management, and human oversight—not just technical monitoring.

🚀 What's Next for Inktrace

Vision: Become the comprehensive security platform enabling safe enterprise agent adoption across Australia's regulated sectors.

  • Phase 1: Proven multi-agent monitoring with Australian compliance ✅
  • Phase 2: ML-based behavioral analysis and advanced threat detection
  • Phase 3: Enterprise integrations (SIEM, SSO, incident response) with upcoming agent ecosystems
  • Phase 4: Agent marketplace security and certification platform

The Reality: Australian enterprises are more likely be buying AI agents, not building them. But there's no security platform to monitor what purchased agents actually do. We're building the category that will define enterprise AI safety in Australia.

Market Opportunity: $200-400M by 2030 as Australia's regulated industries adopt AI agent ecosystems.

Built With

Share this project:

Updates