Security tooling. Privacy-first infrastructure. Human-AI systems built to last.
mazzeleczzare.com · ORCID · Get in touch
I make security practical and AI systems trustworthy — for independent builders, researchers, and organizations that can't afford to get it wrong.
My work sits at the intersection of operational security, privacy infrastructure, and human-AI alignment. Every project here is built against the same constraint: it has to work for people who aren't security engineers.
| Project | What it does |
|---|---|
| secure-pride | Privacy-first cybersecurity tools and standards for LGBTQ+ communities and high-risk groups, built for adversarial conditions |
| secure-container-template | Hardened Python container template — non-root runtime, SBOM, Docker Scout, security-gated CI |
| github-hardening-skills | Idempotent, transferable skills for hardening GitHub repositories — branch protection, Dependabot, CI |
| macos-privacy | Ready-to-deploy, privacy-first macOS terminal + developer environment bootstrap |
| Project | What it does |
|---|---|
| praxis-aegis | Trust-layer for agentic AI: policy guardrails and signing-aware controls for AI tool use |
| stele | Directive compiler and integrity telemetry — governance-as-code for AI-assisted work |
| stratum | Epistemic decision ledger: append-only event log with evidence-gated trust |
| context-synapse | Experimental engine for modeling how humans and AI systems negotiate context |
| Project | What it does |
|---|---|
| meridian | Privacy-first, local-first calendar for Apple platforms — Tailscale-only peer sync, no cloud |
| lockr | Privacy-first dating and social app for the LGBTQ+ community — E2E encryption, safety-first design |
| daedalus | Identity & context switcher — one command switches VPN, terminal, browser, and filesystem context |
| mazze-leczzare-blog | Editorial platform for essays and privacy writing — Astro, MDX, Cloudflare Pages |
- Threat-informed security tooling — reducing real-world risk, not just surface-area compliance
- Privacy-preserving infrastructure — for builders and organizations with something worth protecting
- Legible AI systems — human–AI collaboration that stays calm, auditable, and operationally sound
- Accessible design — correctness and usability aren't in conflict; I refuse to treat them that way
Essays, technical notes, and research on privacy infrastructure, human–AI collaboration, and open-source systems thinking.
→ mazzeleczzare.com · ORCID 0009-0005-9661-4780
Software should be secure, humane, understandable, and accessible to people who aren't security engineers.
That's not a nice-to-have. That's the constraint every project here is built against.
Open to collaboration on privacy tooling, secure infrastructure, and human–AI systems — especially with researchers, independent builders, and mission-driven organizations.
Open an issue, or reach out directly.