Same notes as Writing — how each project was framed, then how it was built.
How v3 is structured — input → analysis → SAFE / SUSPICIOUS / DANGEROUS → next steps, with Postgres and Redis.
Why everyday people need SAFE / SUSPICIOUS / DANGEROUS — not a SOC dashboard — in the moment before they click.
How the toolkit was built — WebSocket progress, check suite, scoring engine, plain-English React report.
The product idea — paste a domain, get a transparent score and plain-English fixes, not a CVE dump.
How the site was shipped — page structure, brand voice, attack-vector animation, Vercel deploy.
Sole designer/developer brief — brand voice, service clarity, and a threat map that earns its place on the homepage.
How the system is structured — Windows agent, collectors, detection and policy engines, React over Node and PostgreSQL.
Why the DLP platform started as architecture — insider leakage, USB, clipboard — and what I refused to fake.
How the lab was put together — payload fixtures, Windows observables, and mitigations next to every demo.
The idea before the Digispark — physical access still wins, and defenders need a safe way to see the chain.