Skip to content
View heinrihs-s's full-sized avatar

Block or report heinrihs-s

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
heinrihs-s/README.md
Heinrihs Skrodelis — security research, useful tools. A control-room single-line diagram with feeders for SCADA/ICS research, AI privacy, agent safety and assistant displays.

I'm a cybersecurity researcher and builder in Latvia. I build privacy tools for AI, little interfaces with personality, and security research for industrial systems.

Website & CV · X / @heinrihs_s · LinkedIn · securIT · Play with a dot →

What I'm building

Send context to the model. Keep identities local.

A self-hosted privacy gateway for AI workflows. Detect sensitive values, apply policy, send scoped tokens to the model, and restore approved values in the reply. Some requests belong on a local model; others should be blocked before they leave.

Explore the code · Try the local playground · How it works

Watch what the model actually sees

Cloakspan terminal demo showing tokenization, policy decisions and approved restoration

The demo uses synthetic data and a deterministic mock provider. Cloakspan is an alpha for local evaluation and synthetic-data pilots.

Give your assistant a little desk space.

An open-source MCP and Home Assistant bridge for assistant displays, including e-ink. Four switchable companions, animated thinking states, and readable replies.

Try the interactive demo → · Build your own · Explore the code

Meet the dots: Artist, Curious, Bookish & Cool

Dots on Paper — Artist, Curious, Bookish and Cool

Open the demo, switch companions, pick an example and replay the thinking state. A browser preview of the idea; the repository contains the bridge and rendering tools.

The research behind the builds

My security work focuses on SCADA/ICS, protocol-aware detection and AI-agent safety. I like small, reproducible artifacts and explanations a tired human can still review.

Disclosure What I found
CVE-2026-62972 · BACnet Stack Out-of-bounds read in the lighting command decoder. Credited as reporter.
OpENer #576 · EtherNet/IP Independent report merged into the response-side buffer overflow investigation, with AFL++ / ASAN reproduction and capacity guards.
  • SCADA Agent SafetyBench — simulation-only evaluation of AI assistants in SCADA / IEC-104 workflows.
  • CTF writeups — challenge notes, including the SANS European Championship 2026.
Open the rest of the lab notebook

Ideas and feedback are welcome in each project's issues. You can also find me on X.

Pinned Loading

  1. quiethours-eu/cloakspan quiethours-eu/cloakspan Public

    Self-hosted privacy firewall for LLM traffic. Detect, pseudonymize, route, and safely restore sensitive data through an OpenAI-compatible API.

    Python 3

  2. dots-on-paper dots-on-paper Public

    A tiny e-ink companion for useful AI replies. Local bridge, MCP, and Home Assistant.

    Python

  3. Scada-Agent-SafetyBench Scada-Agent-SafetyBench Public

    Simulation-only benchmark for AI-agent safety in SCADA/IEC-104 control-room workflows.

    Python

  4. CTF-writeups CTF-writeups Public

    CTF writeups, including SANS European Championship 2026 notes.

    Python