Skip to content
View williamzujkowski's full-sized avatar
💭
Secure infrastructure and LEGO forts -- equally over-engineered.
💭
Secure infrastructure and LEGO forts -- equally over-engineered.

Highlights

  • Pro

Block or report williamzujkowski

Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
williamzujkowski/README.md

Animated terminal showing William's GitHub profile — security engineer, multi-agent AI builder, homelabber, dad-joke connoisseur

This terminal is alive — a scrolling session that refreshes every 6 hours with live weather, live GitHub stats + language breakdown, and a rotating dad joke. The theme rotates daily through 20 palettes.
Built with svg-terminal (48 blocks, zero runtime deps). Set prefers-reduced-motion and you get the static version automatically.


Hey, I'm William

Security Engineer · Multi-Agent AI Builder · Homelabber · Dad Joke Connoisseur

Started breaking and then fixing computers as a kid. Twenty years later:
enterprise vulnerability management across 100k+ assets, then cloud platform security at the federal level.
Now I build multi-agent AI orchestration, run bare-metal Cloud Foundry on a Dell R910 in my basement,
and collect dad jokes at an alarming rate.

📍 /now — what I'm actually working on this week  ·  🛠️ /uses — the homelab + tools  ·  📚 /posts — long-form


🎵 The one most people find me through

If you came here because of live-coding-music-mcp — yes, that one — welcome. It's an MCP server that hands Claude the keys to Strudel.cc for AI-assisted live music coding. It picked up 200+ stars seemingly overnight and I'm still slightly shocked.

The rest of what I build lives below.


What I'm focused on

  • Multi-agent AI governancenexus-agents gives AI coding agents the guardrails I wish I had: adversarial review, drift-detected rules, immutable audit, closed-loop telemetry.
  • Security engineering — Compliance automation, vuln scanning, supply chain. The boring stuff that prevents the exciting stuff. Lately: a Rust vuln-aggregator pulling NVD + EPSS + CISA KEV, and eBPF for container-escape detection.
  • Homelab — Bare-metal BOSH + Cloud Foundry on a Dell R910 (256 GB RAM, 48 threads, Incus/Ubuntu 24.04). Concourse CI, dual Pi-hole, Authentik SSO. Yes, on purpose. Yes, $150/month in power.

Featured projects

Project Why it exists
live-coding-music-mcp Lets Claude drive Strudel.cc for AI-assisted live music coding Stars
standards Battle-tested LLM software-development standards — start any project right in 30 seconds Stars
nexus-agents Governance substrate for AI coding agents — adversarial review, drift-detected rules, immutable audit Stars
cognitive-toolworks Generate cross-platform agent artifacts (SKILL.md, AGENTS.md) using LLM intelligence Stars
svg-terminal Generates ↑ that thing at the top ↑ — 48 blocks, 20 themes, OKLCH WCAG-AAA palettes, deterministic CI builds Stars
mcp-standards-server MCP server for NIST 800-53r5 compliance checking and standards enforcement Stars

Browse the rest at github.com/williamzujkowski?tab=repositories — roughly 80 public repos and one or two I keep meaning to archive.


From the blog

→ Full archive at williamzujkowski.github.io/posts. Auto-refreshes here daily.


Website /now GitHub LinkedIn

Views are my own. The dad jokes are not.

Pinned Loading

  1. live-coding-music-mcp live-coding-music-mcp Public

    A Model Context Protocol (MCP) server that gives Claude direct control over Strudel.cc for AI-assisted music generation and live coding.

    TypeScript 217 28

  2. nexus-substrate/nexus-agents nexus-substrate/nexus-agents Public

    Governance substrate for your AI coding agents — adversarial review, drift-detected rules, immutable audit, closed-loop telemetry

    HTML 15 1

  3. williamzujkowski.github.io williamzujkowski.github.io Public

    Personal website of William Zujkowski, built with Astro and Svelte, styled with Tailwind CSS, and hosted on GitHub Pages.

    Python 1

  4. cognitive-toolworks cognitive-toolworks Public

    Generate cross-platform agent artifacts (SKILL.md, AGENTS.md) using LLM intelligence.

    Python 6

  5. moltdown moltdown Public

    A reproducible, low-click workflow for creating and managing Ubuntu 24.04 Desktop VMs optimized for AI agent work. Run agents with full autonomy, then molt back to a pristine golden image.

    Shell 1

  6. standards standards Public

    LLM Software Development Standards Start any project right in 30 seconds. Battle-tested standards from real production systems.

    Python 17 1