Richard W. DeVaul

Moonshot Engineer · AI/Agent Systems · Space & Propulsion · Mad Scientist
GitHub X / Twitter Mastodon CV (PDF)
Projects Podcast Research About Archive

Projects

Open-source tools and research in AI safety, agentic systems, and beyond.

AI Safety

prompt-lint

Lightweight, dependency-free prompt injection detection for AI agent pipelines. Rule-based pattern matching + statistical n-gram scoring. Detects instruction override, role hijack, exfiltration, tool abuse, and cross-document fragmented injection.

GitHub → Position Paper →
AI Safety

skill-signer

Cryptographic signing and verification for AI agent skills. SSH Ed25519 signatures, MANIFEST.sig.json integrity chains, and allowlist-based execution control. Built for the OpenClaw/ClawHub ecosystem.

GitHub →
CAD/CAM

yapCAD

Open-source agentic CAD/CAM tool — “yet another procedural CAD.” Computational geometry with an intelligence-as-capital approach where skills compound over time rather than consuming tokens repeatedly.

GitHub →
Agent Infrastructure

GLaDOS

A personal AI agent built on OpenClaw — multi-modal, multi-channel assistant with voice interface, automated research briefings, household multi-agent architecture, and proactive monitoring. The system behind the podcast below.

OpenClaw Docs →

Podcast

AI-generated daily briefings on technology, AI, and space.

🎤 GLaDOS Daily Briefing

A daily AI-produced tech briefing covering AI/ML, space & aerospace, agent frameworks, and emerging technology. Two episodes daily: a morning tech roundup and an agent platform deep-dive. Researched, fact-checked, and narrated by AI — because the future should brief itself.

▶ Spotify 📡 RSS Feed

Research Reports

In-depth analysis on emerging technology trends. Updated regularly.

Zero Human Companies & AI-Driven Autonomous Businesses

First published March 3, 2026 · Updated weekly
Read PDF

About

Richard W. DeVaul, Ph.D. is a moonshot engineer at Dark Matter Lab (Relativity Space) in Long Beach, CA, working on space launch, propulsion, and advanced manufacturing.

Previously: co-founder of XNET Mobile (decentralized 4G/5G telecommunications); led innovation and rapid-prototyping teams at Google X and Apple; Ph.D. from the MIT Media Lab in machine learning, wearables, and context-aware human-computer interaction.

Current interests include hybrid propulsion systems, resonant wireless power transfer, agentic engineering (yapCAD), and building AI assistant infrastructure that’s actually useful — with a focus on safety, cryptographic trust, and practical multi-agent architectures.

📧 Reach out via GitHub or X/Twitter.

Archive

Earlier whitepapers and publications.

© 2026 Richard W. DeVaul · GitHub · Legacy site