Blog

Thoughts on technology, cybersecurity, and creative problem-solving

Breaking (and Observing) a Chat Model: My Prompt-Injection Test

A concise write-up of a small experiment: how a model refused a disallowed request, then was socially engineered into producing dangerous output via simple self-attestation.

Read more →

Pentest-Agent — building an LLM-aware red-team assistant

Pentest-Agent is an LLM-enabled penetration-testing orchestrator built to automate reconnaissance and vuln-enumeration workflows while keeping humans squarely in the decision loop. It's plugin-driven, async-first, and pragmatic — it runs `nmap` scans, parses results into structured data, feeds them to an LLM for analysis/planning, and then executes follow-up actions with safety checks. This post covers architecture, key design choices, stumbling blocks, security/ethics, and a roadmap.

Read more →

Building a Web Experience: Creating the Bluedot Website

A technical deep dive into building a full‑stack portfolio using Next.js App Router, Prisma, NextAuth, and GitHub API.

Read more →

Stay Updated

Follow my blog for insights on cybersecurity, development, and creative projects. I share practical tips and behind-the-scenes looks at my work.

Subscribe to Updates