Founder @ DCYFR Labs | Head of AI @ GameShark Labs

github

projects

Drew ✨ • Peerlist

<h3>About Drew Gowan, Cyber Architect</h3>
Drew Gowan is a cyber architect specializing in building autonomous systems that can take real actions in production environments. He focuses on developing robust guardrails to ensure these systems operate safely and reliably. Drew is the Founder at DCYFR Labs and serves as the Head of AI at GameShark Labs.
<ul>
<li>Expertise: Autonomous AI agents, Agentic AI security, AI safety, Multi-agent systems, Local-first inference, Model Context Protocol, Security architecture, and TypeScript.</li>
<li>Core Functionality: Develops autonomous systems with integrated safety mechanisms.</li>
<li>Key Initiatives:
<ul>
<li>@dcyfr/ai: An open-source portable TypeScript runtime for model routing, tool-use, and MCP wiring.</li>
<li>The agentic fleet: A system of daemons running autonomously on a local-first model stack for research, code review, monitoring, and self-healing.</li>
<li>SharkVault: A consumer application built end-to-end by agents, demonstrating autonomous delivery from authentication to content pipeline.</li>
</ul>
</li>
<li>The Loop: Describes the autonomous operational cycle of the agents, including Observe, Triage, Think, Act, and Record states, with built-in safety breaks.</li>
<li>Guardrails: Implements critical safety features such as spend ceilings, sender kill switches, credential isolation, tool sandboxing, metacognition breakers, and self-heal watchdogs to ensure controlled autonomy.</li>
<li>Services Offered:
<ul>
<li>Assistance with building and deploying the first secure agent.</li>
<li>Implementing governance for existing agent fleets.</li>
<li>Auditing current systems for agent capabilities and risks.</li>
</ul>
</li>
<li>Writing: Shares insights on AI, security, and development through blog posts on topics like AI assistants, agentic AI security, and system architecture.</li>
</ul>

Drew Gowan, Cyber Architect

<h3>About @dcyfr/ai</h3>Portable AI agent harness with plugin architecture for managing multiple AI providers, tracking telemetry, and ensuring quality compliance.This framework is designed to streamline the development of AI-powered applications by providing a robust and extensible architecture. It supports integration with various AI providers, enabling developers to leverage different models and services seamlessly.<h3>Key Features:</h3><ul><li>Plugin Architecture: Extensible system for adding custom agents and validation logic.</li><li>Multi-Provider Support: Integrates with major AI providers like OpenAI, Anthropic, Ollama, and GitHub Copilot.</li><li>Telemetry Tracking: Built-in capabilities to monitor usage, costs, and performance metrics.</li><li>Configuration System: Flexible configuration options via YAML, JSON, or package.json, with environment variable overrides.</li><li>Validation Harness: Includes quality gates and validation agents to ensure code and data quality.</li><li>Memory Management: Persistent and working memory options for stateful AI agents.</li><li>CLI Interface: A command-line interface for easy management, configuration, and validation.</li></ul>

DCYFR AI

DCYFR Labs is a virtual partnership dedicated to building secure, innovative solutions for the modern web. We focus on cyber architecture and design, helping organizations navigate the complex landscape of emerging technologies while maintaining robust security postures.<h3>Our Philosophy</h3>“Security isn't about saying no; it's about enabling innovation with confidence.”We believe in building resilient systems that empower teams to move fast without compromising on security. Our approach is practical, human-centric, and focused on real-world constraints rather than theoretical perfection.<h3>What We Focus On</h3>Development &amp; ArchitectureProduction-grade architecture decisions, server-first rendering, and building systems that scale. We share real experiences from real projects, complete with trade-offs and lessons learned.<a target="_blank" rel="noopener noreferrer" class="break-all link-break text-gray-500 hover:underline ml-1 underline underline-offset-4 decoration-skip-ink-none will-change-auto transition-theme active:opacity-70" href="https://www.dcyfr.ai/blog?category=Web">Explore dev articles</a>Cybersecurity &amp; DefenseDefense-in-depth strategies, practical security hardening, and analysis of real threats and defenses. Security as an enabler, not a bottleneck.<a target="_blank" rel="noopener noreferrer" class="break-all link-break text-gray-500 hover:underline ml-1 underline underline-offset-4 decoration-skip-ink-none will-change-auto transition-theme active:opacity-70" href="https://www.dcyfr.ai/blog?category=DevSecOps">Explore security articles</a>AI &amp; AutomationAugmentation over replacement. We explore practical AI workflows, human-centric automation, and how AI can enhance (not replace) human expertise and creativity.<a target="_blank" rel="noopener noreferrer" class="break-all link-break text-gray-500 hover:underline ml-1 underline underline-offset-4 decoration-skip-ink-none will-change-auto transition-theme active:opacity-70" href="https://www.dcyfr.ai/blog?category=ai">Explore AI articles</a><h3>How We Work</h3><ul><li>Real projects, real constraints: We share from actual experience, not ivory tower theory</li><li>Practical over academic: Solutions that work in production matter beyond perfect architecture diagrams</li><li>Transparent trade-offs: Every decision has costs—we discuss them openly</li><li>Open-source contributions: We give back to the community that enables our work</li><li>Continuous learning: Technology evolves, and so do we</li></ul>Through our <a target="_blank" rel="noopener noreferrer" class="break-all link-break text-gray-500 hover:underline underline underline-offset-4 decoration-skip-ink-none will-change-auto transition-theme active:opacity-70" href="https://www.dcyfr.ai/blog">blog</a> and <a target="_blank" rel="noopener noreferrer" class="break-all link-break text-gray-500 hover:underline underline underline-offset-4 decoration-skip-ink-none will-change-auto transition-theme active:opacity-70" href="https://www.dcyfr.ai/work">projects</a>, we share knowledge, insights, and practical guidance for building secure, scalable systems in the age of AI and cloud-native infrastructure.