Skip to content
View hiagokinlevi's full-sized avatar

Block or report hiagokinlevi

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

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

Report abuse
hiagokinlevi/README.md

Hiago Kin Levi

Defensive cybersecurity · open education · governed AI research

I have worked in cybersecurity since 2007. I build defensive tools, study how AI systems can operate within verifiable limits, and publish free learning material for people who want to protect systems and the people who depend on them.

Start learning

OpenCyber Academy

OpenCyber Academy is a free, public cybersecurity learning library. You can study without creating an account. Choose a subject or a formation level, then follow a course's guide, lessons, examples, exercises and worked answers.

Open the academy · Browse by subject · Browse by level · Read the source on GitHub

Start with a published chapter: English — Cybersecurity foundations · Português — Segurança da informação · Español — Seguridad de la información

Each published chapter leads to teaching, a worked example, exercises and answers. For the full route, follow the study path; for a short introduction to send someone, use Share a lesson. The book standard explains how published teaching is reviewed and how it differs from a course plan.

Book

Confronto Cibernético examines the response to advanced persistent threats (APTs). It connects incident investigation, digital forensics, containment and tactical intelligence, and introduces a coordinated defensive “black team” approach. Explore the book and its editions.

Research

These are research projects with their own assumptions and limitations, separate from the academy's course books.

Defensive engineering

Selected public work includes incident-response playbooks, AI security guardrails and cloud posture assessment. See the full project index for more tools and references.

My work favors authorized defensive use, reproducible evidence, clear limitations and source attribution. Original OpenCyber Academy teaching is shared under CC0 1.0 where the rights statement applies; linked sources and other projects keep their own terms.

Connect

Personal website · OpenCyber Academy · LinkedIn

Pinned Loading

  1. OpenCyberAcademy OpenCyberAcademy Public

    OpenCyber Academy — free public cybersecurity courses, books, labs, and AI learning resources in Markdown.

    Python 2

  2. DAI-Decripte-Artificial-Intelligence-LLM DAI-Decripte-Artificial-Intelligence-LLM Public

    DAI: A Domain-Specific LLM with Model-Excluded Authority, Capability-Disjoint Verification, and Recomputable Receipts. Research paper by Hiago Kin Levi, Decripte. Topics: cybersecurity LLMs, inform…

    HTML

  3. ADAI-Autonomous-Defensive-Artificial-Intelligence ADAI-Autonomous-Defensive-Artificial-Intelligence Public

    Autonomous Defensive AI: An Envelope-and-Evidence Framework for Governed LLMs Acting on Production Infrastructure. Research paper by Hiago Kin Levi, Decripte. Topics: autonomous defensive AI, AI se…

    HTML

  4. ir-playbooks-automation ir-playbooks-automation Public

    Incident response playbooks, automated triage workflows, and containment automation for SOC teams and blue teamers

    Python 3

  5. ai-security-guardrails ai-security-guardrails Public

    Security guardrails, input/output validation, and governance controls for LLM applications, agents, and RAG systems

    Python 3

  6. cloud-posture-watch cloud-posture-watch Public

    Cloud security posture assessment for AWS, Azure, and GCP — exposure analysis, logging gaps, and configuration drift detection

    Python 3