Securing AI agents before they press send. 20 years in security operations, now building the tools I wish I had in the SOC.
I build AegisGate — a suite of 5 open-source AI security tools that protect every layer where AI agents interact with systems: HTTP APIs, MCP tool calls, A2A agent communication, browser extensions, IDE plugins, and CI/CD pipelines.
Everything is Apache 2.0, no telemetry, air-gapped capable. Built for environments where security can't phone home.
| Metric | Value |
|---|---|
| Products | 5 (Platform, Rampart, Lens, Scanner, MCP Server) |
| Lines of Go | ~830,000 |
| Tests | 16,400+ (100% detection on real-world payloads, 0% false positives) |
| Detection Patterns | 1,248 (regex + MITRE ATLAS + CharCNN-BiLSTM ML) |
| Compliance Frameworks | 31 (2,043 controls, 1,457 automated) |
| ML Model | CharCNN-BiLSTM, 1.6M params, 100/100 evasion resistance |
| Provisional Patents | 8 (AI agent identity, multi-protocol interception, compliance mapping, three-layer detection, response scanning, static analysis, federated IOC, agent lifecycle) |
| Published Research | SSRN — AegisGate: A Multi-Layered Security Architecture for AI Agent Communication |
| Product | What It Does | Stack |
|---|---|---|
| Platform | Enterprise AI security gateway — reverse proxy with regex + ATLAS + ML detection, 31 compliance frameworks, SIEM integration | Go, gRPC, PostgreSQL, ONNX |
| Rampart | IDE security — VS Code, Cursor, JetBrains plugin that scans prompts before they reach the LLM | Go, LSP |
| Lens | Browser extension — Chrome/Firefox, scans ChatGPT/Claude/Gemini traffic in real-time | JavaScript, WebExtensions |
| Scanner | CLI/CI scanner — static analysis of AI artifacts, SARIF output, GitHub Action | Go, SARIF |
| MCP Server | MCP guardrails — tool-call interception for Model Context Protocol agents | Go, MCP |
- ~20 years in cybersecurity — SOC, incident response, threat intelligence, SIEM engineering
- Security stacks: Splunk, QRadar, XSIAM, Cortex XDR, Cribl, Recorded Future, CyberArk
- Compliance: HIPAA, PCI-DSS, SOC 2, NIST CSF, NIST AI RMF, ISO 27001, ISO 42001, EU AI Act, FISMA, CSA AICM
- Cloud: AWS (GovCloud), Azure (GCC-Moderate, GCC-HIGH), GCP, Docker, Kubernetes, Helm
- Languages: Go, Python, with CI/CD, DevSecOps, SBOM, Cosign, Sigstore
Palo Alto Networks (PCSAE, PMXdS, PMXdR, PMXrC) · Cribl (Level 2 Admin) · Arista (Edge Threat Management) · Recorded Future (Intelligence Fundamentals)
If my open-source work helps your team, consider sponsoring me on GitHub. Every dollar goes to development, CI/CD, and keeping all five products free.
Looking for senior security engineering, AI security, or security architect roles. Open to remote, hybrid, or on-site in the Midwest.